In [1]:
import sys
print sys.version
from joblib import Parallel, delayed
import multiprocessing
nCores = multiprocessing.cpu_count() - 2 # Allow other apps to run
print 'nCores: %d' % (nCores)
from datetime import datetime, time
print 'now: %s' % str(datetime.now())
2.7.11 (default, Jan 28 2016, 14:07:46) [GCC 4.2.1 Compatible Apple LLVM 7.0.2 (clang-700.1.81)] nCores: 14 now: 2016-05-23 08:15:24.694001
In [2]:
import matplotlib.pyplot as plt
%matplotlib inline
from IPython.display import display, Image
import rpy2.robjects as robjects
from rpy2.robjects.packages import importr
from rpy2.robjects.lib import grid
from rpy2.robjects.lib import ggplot2
import rpy2.robjects.pandas2ri
import numpy as np
np.set_printoptions(precision=4, suppress=True)
import os
import pandas as pd
from scipy import ndimage
from six.moves.urllib.request import urlretrieve
from six.moves import cPickle as pickle
import tensorflow as tf
print 'tf.__version__:%s' % str(tf.__version__)
%run img_utils.py
tf.__version__:0.8.0
/usr/local/lib/python2.7/site-packages/rpy2/robjects/lib/ggplot2.py:59: UserWarning: This was designed againt ggplot2 version 2.0.0 but you have 2.1.0
warnings.warn('This was designed againt ggplot2 version %s but you have %s' % (TARGET_VERSION, ggplot2.__version__))
Analytics Specs¶
This Project¶
In [3]:
%run img_glbSpec_SFDD_ImgSz_64.py
imported img_glbSpec_SFDD_Img_Sz_64.py
In [4]:
# print '\nglbDataFile: %s' % (glbDataFile)
print '\nglbRspClass: %s' % (glbRspClass)
print 'glbRspClassN: %d' % (glbRspClassN)
print 'glbRspClassDesc: '; print(glbRspClassDesc)
print '\nglbImg:'; print(glbImg)
print '\nglbTfwVarSeed: %d' % (glbTfwVarSeed)
print '\nglbPickleFile: %s' % (glbPickleFile)
glbRspClass: ['c0', 'c1', 'c2', 'c3', 'c4', 'c5', 'c6', 'c7', 'c8', 'c9']
glbRspClassN: 10
glbRspClassDesc:
{'c9': 'talking to passenger', 'c8': 'hair and makeup', 'c3': 'texting - left', 'c2': 'talking on the phone - right', 'c1': 'texting - right', 'c0': 'normal driving', 'c7': 'reaching behind', 'c6': 'drinking', 'c5': 'operating the radio', 'c4': 'talking on the phone - left'}
glbImg:
{'color': False, 'crop': {'x': (80, 560)}, 'shape': (480, 640, 3), 'pxlDepth': 255.0, 'center_scale': True, 'size': 64}
glbTfwVarSeed: 131
glbPickleFile: {'models': 'data/img_M_SFDD_ImgSz_64.pickle', 'data': 'data/img_D_SFDD_ImgSz_64.pickle'}
Import Data¶
This Project¶
In [17]:
%run img_utils.py
glbObsFitIdn, glbObsFitFtr, glbObsFitRsp, \
glbObsVldIdn, glbObsVldFtr, glbObsVldRsp, \
glbObsNewIdn, glbObsNewFtr, glbObsNewRsp, \
sbtNewCorDf, \
_ = myimportDbs(glbPickleFile['data'])
glbObsTrnIdn = glbObsFitIdn + glbObsVldIdn
glbObsTrnFtr = np.vstack((glbObsFitFtr, glbObsVldFtr))
glbObsTrnRsp = np.concatenate((glbObsFitRsp, glbObsVldRsp))
print('Fit pickled set:',
len(glbObsFitIdn), glbObsFitFtr.shape, glbObsFitRsp.shape)
print('Vld pickled set:',
len(glbObsVldIdn), glbObsVldFtr.shape, glbObsVldRsp.shape)
print('Trn pickled set:',
len(glbObsTrnIdn), glbObsTrnFtr.shape, glbObsTrnRsp.shape)
print('New pickled set:',
len(glbObsNewIdn), glbObsNewFtr.shape, glbObsNewRsp.shape)
Importing database from data/img_D_SFDD_ImgSz_64.pickle...
('Fit pickled set:', 18077, (18077, 64, 64), (18077,))
('Vld pickled set:', 4347, (4347, 64, 64), (4347,))
('Trn pickled set:', 22424, (22424, 64, 64), (22424,))
('New pickled set:', 79726, (79726, 64, 64), (79726,))
First reload the data we generated in 1_notmist.ipynb.
In [7]:
# pickle_file = 'data/notMNIST.pickle'
# with open(pickle_file, 'rb') as f:
# save = pickle.load(f)
# glbXFit = save['glbXFit']
# glbYFit = save['glbYFit']
# glbXVld = save['glbXVld']
# glbYVld = save['glbYVld']
# glbXNew = save['glbXNew']
# glbYNew = save['glbYNew']
# del save # hint to help gc free up memory
# print('Training set', glbXFit.shape, glbYFit.shape)
# print('Validation set', glbXVld.shape, glbYVld.shape)
# print('Test set', glbXNew.shape, glbYNew.shape)
Reformat into a shape that's more adapted to the models we're going to train:
- data as a flat matrix,
- labels as float 1-hot encodings.
In [18]:
def lclreformatData(I, X, Y):
X = X.reshape((-1, glbImg['size'] * glbImg['size'])).astype(np.float32)
# Map 0 to [1.0, 0.0, 0.0 ...], 1 to [0.0, 1.0, 0.0 ...]
Y = (np.arange(glbRspClassN) == Y[:,None]).astype(np.float32)
return I, X, Y
glbITrn, glbXTrn, glbYTrn = lclreformatData(
glbObsTrnIdn, glbObsTrnFtr, glbObsTrnRsp)
glbIFit, glbXFit, glbYFit = lclreformatData(
glbObsFitIdn, glbObsFitFtr, glbObsFitRsp)
glbIVld, glbXVld, glbYVld = lclreformatData(
glbObsVldIdn, glbObsVldFtr, glbObsVldRsp)
glbINew, glbXNew, glbYNew = lclreformatData(
glbObsNewIdn, glbObsNewFtr, glbObsNewRsp)
print('Trn reshaped set:', len(glbITrn), glbXTrn.shape, glbYTrn.shape)
print('Fit reshaped set:', len(glbIFit), glbXFit.shape, glbYFit.shape)
print('Vld reshaped set:', len(glbIVld), glbXVld.shape, glbYVld.shape)
print('New reshaped set:', len(glbINew), glbXNew.shape, glbYNew.shape)
('Trn reshaped set:', 22424, (22424, 4096), (22424, 10))
('Fit reshaped set:', 18077, (18077, 4096), (18077, 10))
('Vld reshaped set:', 4347, (4347, 4096), (4347, 10))
('New reshaped set:', 79726, (79726, 4096), (79726, 10))
In [19]:
# Check how much incremental memory is used for Fit obs
del glbObsFitIdn, glbObsFitFtr, glbObsFitRsp
del glbIFit, glbXFit, glbYFit
# Check how much incremental memory is used for Trn obs
# del glbObsTrnIdn, glbObsTrnFtr, glbObsTrnRsp
# del glbITrn, glbXTrn, glbYTrn
In [20]:
print glbObsFitFtr.shape
# print glbObsTrnFtr.shape
--------------------------------------------------------------------------- NameError Traceback (most recent call last) <ipython-input-20-d819111688f1> in <module>() ----> 1 print glbObsFitFtr.shape 2 # print glbObsTrnFtr.shape NameError: name 'glbObsFitFtr' is not defined
Fit Logistic Regression (TensorFlow)¶
Let's now switch to stochastic gradient descent training instead, which is much faster. The graph will be similar, except that instead of holding all the training data into a constant node, we create a Placeholder node which will be fed actual data at every call of sesion.run().
In [9]:
%run img_glbSpec_SFDD_ImgSz_64.py
%run img_utils.py
# Add parameter for lrnRateTfw = 0.5 for GradientDescentOptimizer
# With gradient descent training, even this much data is prohibitive.
# Subset the training data for faster turnaround.
def fitMdlLgtRgrSGDTfw(lclXFit, lclYFit,
nObsFit = 100, nObsBtc = 16,
rotatePby = 0.1, rotateMaxAgl = 5,
nStepsTfw = 10, lrnRateTfw = 0.1,
visualize = False, newObs = False, verbose = False):
from scipy.ndimage import rotate
from sklearn import metrics as skl_metrics
prtStr = '\nLogistic Regression (TensorFlow): ' + \
"nObsFit:%5d; nObsBtc:%5d; " + \
"rotatePby: %.4f; rotateMaxAgl: %3d; " + \
"nStepsTfw:%5d; lrnRateTfw:%.4f "
print(prtStr % ( \
nObsFit, nObsBtc, \
rotatePby, rotateMaxAgl, \
nStepsTfw, lrnRateTfw))
print(" visualize: %s; newObs: %s; verbose: %s" % ( \
visualize, newObs, verbose))
# Ensure each Fit obs used at least once during training
if (nStepsTfw * nObsBtc < nObsFit):
print " nStepsTfw * nObsBtc < nObsFit: %5d < %5d" % (\
nStepsTfw * nObsBtc, nObsFit)
nStepsTfw = nObsFit / nObsBtc + 1
print " overriding nStepsTfw: %5d" % (nStepsTfw)
startTm = datetime.now()
mdlDf = pd.DataFrame({'id': 'LgtRgr.SGD.tfw',
'nObsFit': [nObsFit],
'nObsBtc': [nObsBtc],
'rotatePby': [rotatePby],
'rotateMaxAgl': [rotateMaxAgl],
'nStepsTfw': [nStepsTfw],
'lrnRateTfw': [lrnRateTfw]
})
graph = tf.Graph()
with graph.as_default():
# Input data.
# The training data, we use a placeholder that will be fed
# at run time with a training minibatch.
# The validation data into constants that
# are attached to the graph.
# The tests data is loaded by batch thru a placeholder
# tfwXFit = tf.constant(lclXFit[:nObsFit, :])
# tfwYFit = tf.constant(lclYFit[:nObsFit])
tfwXFit = tf.placeholder(tf.float32,
shape = (nObsBtc, lclXFit.shape[1]))
tfwYFit = tf.placeholder(tf.float32,
shape = (nObsBtc, lclYFit.shape[1]))
tfwXVld = tf.constant(glbXVld)
tfwYVld = tf.constant(glbYVld)
tfwXNew = tf.placeholder(tf.float32,
shape = (glbImg['size'], lclXFit.shape[1]))
tfwYNew = tf.placeholder(tf.float32,
shape = (glbImg['size'], lclYFit.shape[1]))
# Variables.
tf.set_random_seed(glbTfwVarSeed)
# These are the parameters that we are going to be training.
# The weight matrix will be initialized using random valued
# following a (truncated) normal distribution.
# The bias vector get initialized to zero.
tfwW = tf.Variable(
tf.truncated_normal([glbImg['size'] * glbImg['size'],
glbRspClassN]),
name = 'tfwW')
tfwB = tf.Variable(tf.zeros([glbRspClassN]), name = 'tfwB')
if (verbose):
print(' tfwW:', tfwW.initialized_value())
print(' tfwB:', tfwB.initialized_value())
# print 'lblIx:%2d:%s'% \
# (np.vectorize("%.4e".__mod__)(tfwW.value()[:5, lblIx]))
# Training computation.
# We multiply the inputs with the weight matrix, and add bias.
# We compute the softmax and cross-entropy (it's one operation in
# TensorFlow, because it's very common, and it can be optimized).
# We take the average of this cross-entropy across all training
# examples: that's our loss.
logits = tf.matmul(tfwXFit, tfwW) + tfwB
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(logits, tfwYFit))
# Optimizer.
# We are going to find the minimum of this loss using
# gradient descent.
optimizer = (tf.train
.GradientDescentOptimizer(tf.to_float(lrnRateTfw))
.minimize(loss))
# Predictions for the training, validation, and test data.
# These are not part of training, but merely here so that we can
# report accuracy_score figures as we train.
tfwYFitPby = tf.nn.softmax(logits)
tfwYVldPby = tf.nn.softmax(tf.matmul(tfwXVld, tfwW) + tfwB)
tfwYNewPby = tf.nn.softmax(tf.matmul(tfwXNew, tfwW) + tfwB)
def accuracy_score(predictions, labels):
return (1.0 * np.sum(np.argmax(predictions, 1) ==
np.argmax(labels, 1))
/ predictions.shape[0])
tf.set_random_seed(glbTfwVarSeed)
# For image rotation in feed_dict. Separate seed to ensure
# deterministic performance by tf session irrespective of
# rotation randomization
np.random.seed(glbNPySeed)
with tf.Session(graph=graph) as session:
# This is a one-time operation which ensures the parameters get
# initialized as we described in the graph:
# random tfwW for the matrix, zeros for the tfwB.
tf.initialize_all_variables().run()
if verbose:
print(' Initialized')
# Create a dummy feed for test data & occlusion visualization
# btcNewDct = {tfwXNew: glbXNew[:glbImg['size'], :]}
for step in range(int(nStepsTfw)):
# Pick an offset within the training data, which has been
# randomized.
# Note: we could use better randomization across epochs.
offset = (step * nObsBtc) % (nObsFit - nObsBtc)
# Generate a minibatch (w/ or w/o rotation)
if (np.random.rand() > rotatePby):
btcXFit = lclXFit[offset:(offset + nObsBtc), :]
else:
rawXFit = np.reshape(lclXFit[offset:(offset + nObsBtc), :],
(nObsBtc, glbImg['size'], glbImg['size']))
rttXFit = np.zeros_like(rawXFit)
angle = (np.random.rand() - 0.5) * 2 * rotateMaxAgl
if (verbose):
prtStr = ' step %5d(%5d secs): Minibatch rotation:' + \
"angle: %.4f"
print(prtStr % \
(step, thsDrn, \
angle))
# print ' rawXFit.shape: %s' % (str(rawXFit.shape))
for ix in xrange(rawXFit.shape[0]):
# if (ix % 10 == 0):
# print ' rawXFit[ix, :, :].shape: %s' % \
# (str(rawXFit[ix, :, :].shape))
rttXFit[ix, :, :] = rotate(rawXFit[ix, :, :], angle,
mode = 'nearest', reshape = False)
btcXFit = np.reshape(rttXFit,
(nObsBtc, glbImg['size'] * glbImg['size']))
btcYFit = lclYFit[offset:(offset + nObsBtc), :]
# Prepare a dictionary telling the session where to feed the
# minibatch. The key of the dictionary is the placeholder node
# of the graph to be fed, and the value is the numpy array to
# feed to it.
feed_dict = {tfwXFit: btcXFit, tfwYFit: btcYFit,
tfwXNew: glbXNew[:glbImg['size'], :]}
# Run the computations.
# We tell .run() that we want to run the optimizer,
# and get the loss value and the training predictions returned
# as numpy arrays.
_, l, predictions = \
session.run([optimizer, loss, tfwYFitPby],
feed_dict = feed_dict)
if mydspVerboseTrigger(step):
thsDrn = (datetime.now() - startTm).seconds
if (thsDrn > 100):
prtStr = ' step %5d(%5d secs): Minibatch ' + \
"accuracy: %.4f; logloss: %.4f"
print(prtStr % \
(step, thsDrn,
accuracy_score(predictions, btcYFit), l))
# accFit = accuracy_score(tfwYFitPby.eval()[:nObsFit, :],
# tfwYFit.eval() [:nObsFit, :])
# logLossFit = \
# skl_metrics.log_loss(tfwYFit.eval() [:nObsFit, :],
# tfwYFitPby.eval()[:nObsFit, :])
# if verbose:
# print '\n Fit accuracy:%0.4f' % (accFit)
# print ' Fit logLoss:%0.4f' % (logLossFit)
# print metrics.confusion_matrix(glbYFit[0:nObsFit],
# lclYFitPdn)
# Calling .eval() on tfwObsVldPred is basically like calling run(),
# but just to get that one numpy array.
# Note that it recomputes all its graph dependencies.
lclYVldPby = tfwYVldPby.eval()
lclYVldPdn = np.argmax(lclYVldPby, 1)
mdlDf['accVld'] = accVld = accuracy_score(lclYVldPby, glbYVld)
cnfYVld = skl_metrics.confusion_matrix(glbObsVldRsp, lclYVldPdn)
accYVldCls = cnfYVld.diagonal() * 1.0 / cnfYVld.sum(axis = 1)
mdlDf['accVldCls'] = None
mdlDf.set_value(0, 'accVldCls', {'accCls' : accYVldCls})
mdlDf['logLossVld'] = logLossVld = skl_metrics.log_loss(
glbYVld, lclYVldPby)
logLossVldCls = mygetMetricLogLoss(glbYVld, lclYVldPby,
returnTyp = 'class')
mdlDf['logLossVldCls'] = None
mdlDf.set_value(0, 'logLossVldCls',
{'logLossCls' : logLossVldCls})
if verbose:
print '\n Vld accuracy:%0.4f' % (accVld)
print accYVldCls
print cnfYVld
yLbl = [glbRspClassDesc[glbRspClass[ix]] + ':' + \
glbRspClass[ix] + ':actl' \
for ix in xrange(glbRspClassN)]
xLbl = ['pred:' + glbRspClass[ix] + ':' + \
glbRspClassDesc[glbRspClass[ix]] \
for ix in xrange(glbRspClassN)]
# print labels
plt.matshow(cnfYVld, cmap='Reds', interpolation='none')
plt.yticks(np.arange(10), yLbl)
plt.xticks(np.arange(10), xLbl, rotation=90);
plt.show()
print '\n Vld logLoss:%0.4f' % (logLossVld)
print logLossVldCls
if visualize:
mydisplayImagePredictions(session, tfwW.eval(),
glbIVld, glbObsVldFtr, glbObsVldRsp, lclYVldPby,
glbRspClass, glbRspClassDesc, imgVisualFn = None,
tfwXOcc = tfwXNew, tfwYOccPby = tfwYNewPby)
if newObs:
print " predicting %5d new obs..." % (glbYNew.shape[0])
lclYNewPby = np.zeros((glbYNew.shape[0],
tfwYFitPby.get_shape().as_list()[1]))
lclYNewPby[:, :] = -1.0
btcSz = tfwXNew.get_shape().as_list()[0]
for obsIx in xrange(0, glbYNew.shape[0], btcSz):
if mydspVerboseTrigger(obsIx) and \
(datetime.now() - startTm).seconds > 60:
print " @%5d secs: obsIx: %5d" % \
((datetime.now() - startTm).seconds, obsIx)
obsEnd = obsIx + btcSz
if obsEnd > lclYNewPby.shape[0]:
obsEnd = lclYNewPby.shape[0]
btcYNewPby = session.run(tfwYNewPby,
feed_dict = {tfwXNew: glbXNew[obsIx:obsEnd, :] \
if obsEnd != lclYNewPby.shape[0] \
else np.vstack((glbXNew[obsIx:obsEnd, :],
glbXNew[0:((obsIx + btcSz) % obsEnd), :]))
})
lclYNewPby[obsIx:obsEnd, :] = btcYNewPby[:, :] \
if obsEnd != lclYNewPby.shape[0] \
else btcYNewPby[:(obsEnd - obsIx), :]
assert (lclYNewPby[:, :] != -1.0).all(), \
'some cells in lclYNewPby == -1.0'
# lclYNewPdn = tfwYNew.eval()
# lclYNewPby = tfwYNewPby.eval()
lclYNewPdn = np.argmax(lclYNewPby, 1)
#if (tfwYNew.eval() > -1).any():
if (len(np.unique(glbYNew, return_counts = True)[0]) > 1):
mdlDf['accNew'] = accNew = accuracy_score(lclYNewPby,
glbYNew)
mdlDf['logLossNew'] = logLossNew = skl_metrics.log_loss(
glbYNew, lclYNewPby)
if verbose:
print '\n New accuracy:%0.4f' % (accNew)
print ' New logLoss:%0.4f' % (logLossNew)
print skl_metrics.confusion_matrix(glbObsNewRsp,
lclYNewPdn)
yLbl = [glbRspClassDesc[glbRspClass[ix]] + ':' +
glbRspClass[ix] + ':actl' \
for ix in xrange(glbRspClassN)]
xLbl = ['pred:' + glbRspClass[ix] + ':' + \
glbRspClassDesc[glbRspClass[ix]] \
for ix in xrange(glbRspClassN)]
# print labels
plt.matshow(skl_metrics.confusion_matrix(glbObsNewRsp,
lclYNewPdn),
cmap='Reds', interpolation='none')
plt.yticks(np.arange(10), yLbl)
plt.xticks(np.arange(10), xLbl, rotation=90);
plt.show()
if visualize:
mydisplayImagePredictions(session, tfwW.eval(),
glbINew, glbObsNewFtr, glbObsNewRsp, lclYNewPby,
glbRspClass, glbRspClassDesc, imgVisualFn = None,
tfwXOcc = tfwXNew, tfwYOccPby = tfwYNewPby)
mdlDf['predNew'] = None
mdlDf.set_value(0, 'predNew', {'kntCls' : np.unique(lclYNewPdn,
return_counts = True)})
if verbose:
print '\n New prediction knts:'
print mdlDf['predNew'][0]
# indentation (6 spaces) determines scope of this
# before session.__exit__ & graph.__exit__
mdlDf['model'] = session
mdlDf['duration'] = (datetime.now() - startTm).seconds
print(' duration: %.2d seconds' % (mdlDf['duration'][0]))
if not newObs: lclYNewPby = None
return(mdlDf, lclYVldPby, lclYNewPby)
tmpMdlDf = pd.DataFrame()
# thsMdlDf, thsYVldPby, thsYNewPby = fitMdlLgtRgrSGDTfw(
# glbXFit, glbYFit,
# nObsFit = 100, nObsBtc = 16,
# rotatePby = 0.2, rotateMaxAgl = 10,
# nStepsTfw = 10, lrnRateTfw = 0.5,
# visualize = True, newObs = True, verbose = True)
# tmpMdlDf = tmpMdlDf.append(thsMdlDf)
# To check if model results are deterministic &
# all run options work separately
thsMdlDf, thsYVldPby, thsYNewPby = fitMdlLgtRgrSGDTfw(
glbXFit, glbYFit,
nObsFit = 100, nObsBtc = 16,
rotatePby = 0.2, rotateMaxAgl = 10,
nStepsTfw = 10, lrnRateTfw = 0.5,
visualize = True, newObs = False, verbose = False)
tmpMdlDf = tmpMdlDf.append(thsMdlDf)
thsMdlDf, thsYVldPby, thsYNewPby = fitMdlLgtRgrSGDTfw(
glbXFit, glbYFit,
nObsFit = 100, nObsBtc = 16,
rotatePby = 0.2, rotateMaxAgl = 10,
nStepsTfw = 10, lrnRateTfw = 0.5,
visualize = False, newObs = True, verbose = False)
tmpMdlDf = tmpMdlDf.append(thsMdlDf)
thsMdlDf, thsYVldPby, thsYNewPby = fitMdlLgtRgrSGDTfw(
glbXFit, glbYFit,
nObsFit = 100, nObsBtc = 16,
rotatePby = 0.2, rotateMaxAgl = 10,
nStepsTfw = 10, lrnRateTfw = 0.5,
visualize = False, newObs = False, verbose = True)
tmpMdlDf = tmpMdlDf.append(thsMdlDf)
thsMdlDf, thsYVldPby, thsYNewPby = fitMdlLgtRgrSGDTfw(
glbXFit, glbYFit,
nObsFit = 100, nObsBtc = 16,
rotatePby = 0.2, rotateMaxAgl = 10,
nStepsTfw = 10, lrnRateTfw = 0.5,
visualize = False, newObs = False, verbose = False)
tmpMdlDf = tmpMdlDf.append(thsMdlDf)
print '\ntmpMdlDf: '
print(tmpMdlDf)
imported img_glbSpec_SFDD_Img_Sz_64.py Logistic Regression (TensorFlow): nObsFit: 100; nObsBtc: 16; rotatePby: 0.2000; rotateMaxAgl: 10; nStepsTfw: 10; lrnRateTfw:0.5000 visualize: True; newObs: False; verbose: False max Pby for cls: c0; desc: normal driving; proba: 0.9169; nObs: 1 img_23350.jpg: plot_occlusion:
display_weight:
Proba: [ 0.9169 0. 0. 0. 0.083 0. 0. 0.0001 0. 0. ] min Pby for cls: c0; desc: normal driving; proba: 0.9169; nObs: 1 img_23350.jpg: plot_occlusion:
display_weight:
Proba: [ 0.9169 0. 0. 0. 0.083 0. 0. 0.0001 0. 0. ] next best class: talking on the phone - left max Pby for cls: c2; desc: talking on the phone - right; proba: 1.0000; nObs: 179 img_8759.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c2; desc: talking on the phone - right; proba: 0.4792; nObs: 1 img_21880.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.4792 0.425 0. 0. 0. 0.0957 0. 0. ] next best class: texting - left max Pby for cls: c3; desc: texting - left; proba: 1.0000; nObs: 440 img_29251.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.] min Pby for cls: c3; desc: texting - left; proba: 0.4872; nObs: 1 img_74128.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.4701 0.4872 0. 0. 0. 0.0427 0. 0. ] next best class: talking on the phone - right max Pby for cls: c4; desc: talking on the phone - left; proba: 0.9905; nObs: 1 img_53159.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.0023 0.0072 0.9905 0. 0. 0. 0. 0. ] min Pby for cls: c4; desc: talking on the phone - left; proba: 0.9905; nObs: 1 img_53159.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.0023 0.0072 0.9905 0. 0. 0. 0. 0. ] next best class: texting - left max Pby for cls: c6; desc: drinking; proba: 0.8050; nObs: 1 img_22902.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.018 0.1439 0. 0.0332 0. 0.805 0. 0. 0. ] min Pby for cls: c6; desc: drinking; proba: 0.8050; nObs: 1 img_22902.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.018 0.1439 0. 0.0332 0. 0.805 0. 0. 0. ] next best class: talking on the phone - right max Pby for cls: c7; desc: reaching behind; proba: 1.0000; nObs: 4 img_91808.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.] min Pby for cls: c7; desc: reaching behind; proba: 0.5289; nObs: 1 img_35334.jpg: plot_occlusion:
display_weight:
Proba:
[ 0. 0. 0.0049 0.4662 0. 0. 0. 0.5289 0. 0. ]
next best class: texting - left
duration: 26 seconds
Logistic Regression (TensorFlow): nObsFit: 100; nObsBtc: 16; rotatePby: 0.2000; rotateMaxAgl: 10; nStepsTfw: 10; lrnRateTfw:0.5000
visualize: False; newObs: True; verbose: False
predicting 79726 new obs...
duration: 13 seconds
Logistic Regression (TensorFlow): nObsFit: 100; nObsBtc: 16; rotatePby: 0.2000; rotateMaxAgl: 10; nStepsTfw: 10; lrnRateTfw:0.5000
visualize: False; newObs: False; verbose: True
(' tfwW:', <tf.Tensor 'Identity:0' shape=(4096, 10) dtype=float32>)
(' tfwB:', <tf.Tensor 'Identity_1:0' shape=(10,) dtype=float32>)
Initialized
step 1( 0 secs): Minibatch rotation:angle: 3.9074
step 6( 0 secs): Minibatch rotation:angle: 6.9499
Vld accuracy:0.1049
[ 0. 0. 0.5111 0.5056 0. 0. 0. 0.0029 0. 0. ]
[[ 0 0 308 174 0 0 0 2 0 0]
[ 0 0 158 320 0 0 0 0 0 0]
[ 0 0 231 220 0 0 0 1 0 0]
[ 0 0 219 224 0 0 0 0 0 0]
[ 0 0 149 308 0 0 0 2 0 0]
[ 1 0 76 304 0 0 0 41 0 0]
[ 0 0 135 307 0 0 0 1 0 0]
[ 0 0 151 197 1 0 0 1 0 0]
[ 0 0 199 165 0 0 0 0 0 0]
[ 0 0 262 183 0 0 1 6 0 0]]
Vld logLoss:26.7370
[ 3.8456 3.6906 0.3847 0.3825 3.5034 3.353 3.5066 1.5953 2.8921
3.5831]
duration: 01 seconds
Logistic Regression (TensorFlow): nObsFit: 100; nObsBtc: 16; rotatePby: 0.2000; rotateMaxAgl: 10; nStepsTfw: 10; lrnRateTfw:0.5000
visualize: False; newObs: False; verbose: False
duration: 00 seconds
tmpMdlDf:
accVld accVldCls duration \
0 0.1049 {u'accCls': [0.0, 0.0, 0.511061946903, 0.50564... 26
0 0.1049 {u'accCls': [0.0, 0.0, 0.511061946903, 0.50564... 13
0 0.1049 {u'accCls': [0.0, 0.0, 0.511061946903, 0.50564... 1
0 0.1049 {u'accCls': [0.0, 0.0, 0.511061946903, 0.50564... 0
id logLossVld \
0 LgtRgr.SGD.tfw 26.737001
0 LgtRgr.SGD.tfw 26.737001
0 LgtRgr.SGD.tfw 26.737001
0 LgtRgr.SGD.tfw 26.737001
logLossVldCls lrnRateTfw \
0 {u'logLossCls': [3.84558724986, 3.69061398653,... 0.5
0 {u'logLossCls': [3.84558724986, 3.69061398653,... 0.5
0 {u'logLossCls': [3.84558724986, 3.69061398653,... 0.5
0 {u'logLossCls': [3.84558724986, 3.69061398653,... 0.5
model nObsBtc nObsFit \
0 <tensorflow.python.client.session.Session obje... 16 100
0 <tensorflow.python.client.session.Session obje... 16 100
0 <tensorflow.python.client.session.Session obje... 16 100
0 <tensorflow.python.client.session.Session obje... 16 100
nStepsTfw predNew rotateMaxAgl \
0 10 NaN 10
0 10 {u'kntCls': ([1, 2, 3, 4, 6, 7, 8, 9], [157, 5... 10
0 10 NaN 10
0 10 NaN 10
rotatePby
0 0.2
0 0.2
0 0.2
0 0.2
In [30]:
lr1MdlDf, lr1YVldPby, lr1YNewPby = fitMdlLgtRgrTfw(
glbXFit, glbYFit,
nObsFit = 1000, nStepsTfw = 100, lrnRateTfw = 0.1,
visualize = False, newObs = True, verbose = True)
Logistic Regression (TensorFlow): nObsFit: 1000; nStepsTfw: 100; lrnRateTfw:0.1000
visualize: False; newObs: True; verbose: True
(' tfwW:', <tf.Tensor 'Identity:0' shape=(4096, 10) dtype=float32>)
(' tfwB:', <tf.Tensor 'Identity_1:0' shape=(10,) dtype=float32>)
Initialized
Fit accuracy:0.2470
Fit logLoss:7.2014
Vld accuracy:0.1735
Vld logLoss:12.6001
[[ 73 22 44 15 77 7 94 50 18 84]
[ 33 68 51 31 24 22 90 77 10 72]
[ 14 100 61 9 115 1 28 113 2 9]
[ 2 18 29 136 34 7 37 19 10 151]
[ 5 14 17 137 67 9 58 91 3 58]
[ 11 16 19 41 114 13 14 151 30 13]
[ 9 47 31 55 113 9 45 96 19 19]
[ 1 16 10 14 57 28 45 130 24 25]
[ 21 12 47 36 47 16 50 66 8 61]
[ 2 12 15 8 38 9 56 140 19 153]]
predicting 79726 new obs...
New prediction knts:
{'clsKnt': (array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]), array([ 8786, 11144, 10374, 6453, 5874, 4904, 9130, 6548, 7613, 8900]))}
duration: 19 seconds
In [31]:
lr5MdlDf, lr5YVldPby, lr5YNewPby = fitMdlLgtRgrTfw(
glbXFit, glbYFit,
nObsFit = 1000, nStepsTfw = 100, lrnRateTfw = 0.5,
visualize = False, newObs = True, verbose = True)
Logistic Regression (TensorFlow): nObsFit: 1000; nStepsTfw: 100; lrnRateTfw:0.5000
visualize: False; newObs: True; verbose: True
(' tfwW:', <tf.Tensor 'Identity:0' shape=(4096, 10) dtype=float32>)
(' tfwB:', <tf.Tensor 'Identity_1:0' shape=(10,) dtype=float32>)
Initialized
Fit accuracy:0.4970
Fit logLoss:5.6870
Vld accuracy:0.2227
Vld logLoss:13.3804
[[148 76 77 0 65 0 4 41 4 69]
[ 58 87 100 0 44 4 2 121 9 53]
[ 11 63 120 0 110 0 0 121 0 27]
[ 10 36 83 1 132 3 3 31 0 144]
[ 11 19 86 0 197 5 17 53 0 71]
[ 34 7 66 0 79 35 1 121 38 41]
[ 33 35 135 0 134 2 1 73 5 25]
[ 2 15 39 0 46 10 0 191 22 25]
[ 30 24 105 0 40 6 0 62 32 65]
[ 13 12 87 0 47 6 0 128 3 156]]
predicting 79726 new obs...
New prediction knts:
{'clsKnt': (array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]), array([12150, 13070, 22283, 135, 4653, 4845, 390, 7545, 5352, 9303]))}
duration: 19 seconds
In [108]:
%run img_utils.py
from sklearn import metrics as skl_metrics
def lclaccuracy_score(predictions, labels):
return (1.0 * np.sum(np.argmax(predictions, 1) ==
np.argmax(labels, 1))
/ predictions.shape[0])
print "\nlrnRateTfw: 0.1; accVld: %0.4f" % \
(lclaccuracy_score(lr1YVldPby, glbYVld))
cnf1YVld = skl_metrics.confusion_matrix(glbObsVldRsp,
np.argmax(lr1YVldPby, 1))
print cnf1YVld
# print cnf1YVld.sum(axis = 1)
print cnf1YVld.diagonal() * 1.0 / cnf1YVld.sum(axis = 1)
# print cnf1YVld.sum(axis = 1).sum()
lls1YVld = skl_metrics.log_loss(glbYVld, lr1YVldPby)
print "\nlrnRateTfw: 0.1; logLossVld: %0.4f" % (lls1YVld)
# lls1YVldCls = np.array([skl_metrics.log_loss(
# glbYVld[glbYVld[:, clsIx] == 1, :],
# lr1YVldPby[glbYVld[:, clsIx] == 1, :]) \
# for clsIx in range(glbRspClassN)]) / glbRspClassN
# print "lrnRateTfw: 0.1; logLossVld.skl classSum: %0.4f" % (lls1YVld.sum())
# print lls1YVld
lls1YVldCls = mygetMetricLogLoss(glbYVld, lr1YVldPby, returnTyp = 'class')
print "lrnRateTfw: 0.1; logLossVldCls :"; print lls1YVldCls
# print "diff: %.4e" % (np.abs(lls1YVld - lls1YVldCls.sum()))
assert np.abs(lls1YVld - lls1YVldCls.sum()) < 1e-04, \
"logLoss from skl: %.4f vs myCls %.4f does not match" % \
(lls1YVld, lls1YVldCls.sum())
print "\nlrnRateTfw: 0.5; accVld: %0.4f" % \
(lclaccuracy_score(lr5YVldPby, glbYVld))
cnf5YVld = skl_metrics.confusion_matrix(glbObsVldRsp,
np.argmax(lr5YVldPby, 1))
print cnf5YVld
print cnf5YVld.diagonal() * 1.0 / cnf5YVld.sum(axis = 1)
lls5YVld = skl_metrics.log_loss(glbYVld, lr5YVldPby)
print "\nlrnRateTfw: 0.5; logLossVld: %0.4f" % (lls5YVld)
lls5YVldCls = mygetMetricLogLoss(glbYVld, lr5YVldPby, returnTyp = 'class')
print "lrnRateTfw: 0.5; logLossVldCls :"; print lls5YVldCls
assert np.abs(lls5YVld - lls5YVldCls.sum()) < 1e-04, \
"logLoss from skl: %.4f vs myCls %.4f does not match" % \
(lls5YVld, lls5YVldCls.sum())
lrnRateTfw: 0.1; accVld: 0.1735 [[ 73 22 44 15 77 7 94 50 18 84] [ 33 68 51 31 24 22 90 77 10 72] [ 14 100 61 9 115 1 28 113 2 9] [ 2 18 29 136 34 7 37 19 10 151] [ 5 14 17 137 67 9 58 91 3 58] [ 11 16 19 41 114 13 14 151 30 13] [ 9 47 31 55 113 9 45 96 19 19] [ 1 16 10 14 57 28 45 130 24 25] [ 21 12 47 36 47 16 50 66 8 61] [ 2 12 15 8 38 9 56 140 19 153]] [ 0.1508 0.1423 0.135 0.307 0.146 0.0308 0.1016 0.3714 0.022 0.3385] lrnRateTfw: 0.1; logLossVld: 12.6001 lrnRateTfw: 0.1; logLossVldCls : [ 1.4741 2.0839 1.4873 0.8025 1.0307 1.8989 1.0157 0.5303 1.2532 1.0236] lrnRateTfw: 0.5; accVld: 0.2227 [[148 76 77 0 65 0 4 41 4 69] [ 58 87 100 0 44 4 2 121 9 53] [ 11 63 120 0 110 0 0 121 0 27] [ 10 36 83 1 132 3 3 31 0 144] [ 11 19 86 0 197 5 17 53 0 71] [ 34 7 66 0 79 35 1 121 38 41] [ 33 35 135 0 134 2 1 73 5 25] [ 2 15 39 0 46 10 0 191 22 25] [ 30 24 105 0 40 6 0 62 32 65] [ 13 12 87 0 47 6 0 128 3 156]] [ 0.3058 0.182 0.2655 0.0023 0.4292 0.0829 0.0023 0.5457 0.0879 0.3451] lrnRateTfw: 0.5; logLossVld: 13.3804 lrnRateTfw: 0.5; logLossVldCls : [ 1.0024 1.5107 0.8492 2.5585 0.8238 1.6644 2.5152 0.3775 1.327 0.7518]
In [10]:
# glbMdlDf = None
# glbMdlDf = pd.DataFrame()
try:
with open(glbPickleFile['models'], 'rb') as f:
glbMdlDf = pickle.load(f)
assert isinstance(glbMdlDf, pd.DataFrame), \
'type(glbMdlDf): %s, expecting pd.DataFrame' % \
(str(type(glbMdlDf)))
except IOError, e:
print e
print 'file %s not present or not appropriate' % \
(glbPickleFile['models'])
print glbMdlDf
accVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 1.000000
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.987348
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.381643
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0.375201
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 0.361859
2.0 1 0.357488
0.0 8.0 3.0 0 0.357028
2 0.357028
1 0.357028
10 0.357028
0.1 8.0 3.0 0 0.357028
4.0 1.0 1 0.355648
8.0 2.0 2 0.351737
4 0.348516
16.0 4.0 10 0.346216
0.0 8.0 7.0 10 0.346216
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 0.343685
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 0.338854
1 0.338854
8.0 5.0 10 0.336784
0.5 8.0 4.0 10 0.334484
16.0 4.0 10 0.331723
0.0 8.0 1.0 0 0.331493
1 0.331493
2 0.331493
10 0.331493
0.1 8.0 1.0 0 0.331493
0.0 32.0 1.0 10 0.328042
0.1 8.0 2.0 5 0.325512
0.0 8.0 0.5 10 0.325052
... ...
0.1 4.0 3.0 1 0.226823
16.0 3.0 10 0.219692
0.0 16.0 2.0 2 0.214861
10 0.214861
1 0.214861
0 0.214861
100.0 18077.0 0.1 16.0 1.0 10 0.210950
1000.0 18077.0 0.5 4.0 3.0 10 0.208420
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0.342535
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0.384173
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 0.381643
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0.358638
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 0.319991
0.1 0 0.251208
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 0.210950
1000.0 9000.0 0.1 8.0 3.0 2 0.381643
8000.0 0.1 8.0 3.0 2 0.378882
7000.0 0.1 8.0 3.0 2 0.375431
6000.0 0.1 8.0 3.0 2 0.339545
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0.363699
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 0.320911
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 0.341385
1000.0 0.0 1000.0 -1.0 0 0.333333
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 0.273522
1.0 0 0.255809
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 0.224983
100.0 1000.0 0.1 16.0 1.0 10 0.212790
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0.173453
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0.307108
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 0.224753
accVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'accCls': [0.989669421488, 0.991631799163, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'accCls': [0.504132231405, 0.435146443515, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'accCls': [0.448347107438, 0.44769874477, 0....
2.0 1 {u'accCls': [0.338842975207, 0.347280334728, 0...
0.0 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
2 {u'accCls': [0.460743801653, 0.640167364017, 0...
1 {u'accCls': [0.460743801653, 0.640167364017, 0...
10 {u'accCls': [0.460743801653, 0.640167364017, 0...
0.1 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
4.0 1.0 1 {u'accCls': [0.52479338843, 0.575313807531, 0....
8.0 2.0 2 {u'accCls': [0.146694214876, 0.5, 0.2123893805...
4 {u'accCls': [0.384297520661, 0.297071129707, 0...
16.0 4.0 10 {u'accCls': [0.510330578512, 0.389121338912, 0...
0.0 8.0 7.0 10 {u'accCls': [0.421487603306, 0.410041841004, 0...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'accCls': [0.506198347107, 0.644351464435, 0...
1 {u'accCls': [0.506198347107, 0.644351464435, 0...
8.0 5.0 10 {u'accCls': [0.338842975207, 0.173640167364, 0...
0.5 8.0 4.0 10 {u'accCls': [0.421487603306, 0.558577405858, 0...
16.0 4.0 10 {u'accCls': [0.396694214876, 0.0836820083682, ...
0.0 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
1 {u'accCls': [0.25, 0.246861924686, 0.303097345...
2 {u'accCls': [0.25, 0.246861924686, 0.303097345...
10 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.1 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.0 32.0 1.0 10 {u'accCls': [0.588842975207, 0.133891213389, 0...
0.1 8.0 2.0 5 {u'accCls': [0.276859504132, 0.219665271967, 0...
0.0 8.0 0.5 10 {u'accCls': [0.365702479339, 0.119246861925, 0...
... ...
0.1 4.0 3.0 1 {u'accCls': [0.545454545455, 0.0690376569038, ...
16.0 3.0 10 {u'accCls': [0.528925619835, 0.0523012552301, ...
0.0 16.0 2.0 2 {u'accCls': [0.495867768595, 0.0397489539749, ...
10 {u'accCls': [0.495867768595, 0.0397489539749, ...
1 {u'accCls': [0.495867768595, 0.0397489539749, ...
0 {u'accCls': [0.495867768595, 0.0397489539749, ...
100.0 18077.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 18077.0 0.5 4.0 3.0 10 {u'accCls': [0.297520661157, 0.081589958159, 0...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'accCls': [0.456611570248, 0.380753138075, 0...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'accCls': [0.289256198347, 0.182008368201, 0...
0.1 0 {u'accCls': [0.169421487603, 0.144351464435, 0...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 9000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
8000.0 0.1 8.0 3.0 2 {u'accCls': [0.553719008264, 0.610878661088, 0...
7000.0 0.1 8.0 3.0 2 {u'accCls': [0.431818181818, 0.558577405858, 0...
6000.0 0.1 8.0 3.0 2 {u'accCls': [0.181818181818, 0.671548117155, 0...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'accCls': [0.543388429752, 0.535564853556, 0...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'accCls': [0.646694214876, 0.0564853556485, ...
1.0 0 {u'accCls': [0.423553719008, 0.182008368201, 0...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'accCls': [0.204545454545, 0.0334728033473, ...
100.0 1000.0 0.1 16.0 1.0 10 {u'accCls': [0.136363636364, 0.0230125523013, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'accCls': [0.150826446281, 0.142259414226, 0...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'accCls': [0.0185950413223, 0.133891213389, ...
bestFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 False
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 True
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 False
2.0 1 False
0.0 8.0 3.0 0 False
2 False
1 False
10 False
0.1 8.0 3.0 0 False
4.0 1.0 1 False
8.0 2.0 2 False
4 False
16.0 4.0 10 False
0.0 8.0 7.0 10 False
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 False
1 False
8.0 5.0 10 False
0.5 8.0 4.0 10 False
16.0 4.0 10 False
0.0 8.0 1.0 0 False
1 False
2 False
10 False
0.1 8.0 1.0 0 False
0.0 32.0 1.0 10 False
0.1 8.0 2.0 5 False
0.0 8.0 0.5 10 False
... ...
0.1 4.0 3.0 1 False
16.0 3.0 10 False
0.0 16.0 2.0 2 False
10 False
1 False
0 False
100.0 18077.0 0.1 16.0 1.0 10 False
1000.0 18077.0 0.5 4.0 3.0 10 False
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 False
0.1 0 False
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 False
1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 False
1000.0 0.0 1000.0 -1.0 0 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 False
1.0 0 False
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 False
100.0 1000.0 0.1 16.0 1.0 10 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 False
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 False
duration \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 874
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 743
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 11
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 519
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 9
2.0 1 10
0.0 8.0 3.0 0 8
2 8
1 9
10 10
0.1 8.0 3.0 0 10
4.0 1.0 1 8
8.0 2.0 2 10
4 13
16.0 4.0 10 14
0.0 8.0 7.0 10 9
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 597
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 8
1 8
8.0 5.0 10 9
0.5 8.0 4.0 10 15
16.0 4.0 10 24
0.0 8.0 1.0 0 8
1 8
2 8
10 9
0.1 8.0 1.0 0 13
0.0 32.0 1.0 10 16
0.1 8.0 2.0 5 10
0.0 8.0 0.5 10 9
... ...
0.1 4.0 3.0 1 9
16.0 3.0 10 13
0.0 16.0 2.0 2 10
10 11
1 11
0 11
100.0 18077.0 0.1 16.0 1.0 10 2
1000.0 18077.0 0.5 4.0 3.0 10 12
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 514
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 288
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 317
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 301
0.1 0 295
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 2
1000.0 9000.0 0.1 8.0 3.0 2 9
8000.0 0.1 8.0 3.0 2 9
7000.0 0.1 8.0 3.0 2 9
6000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 150
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 55
1000.0 0.0 1000.0 -1.0 0 28
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 7
1.0 0 5
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 9
100.0 1000.0 0.1 16.0 1.0 10 2
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 5
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 8
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 9
id \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 LgtRgr.SGD.tfw
2.0 1 LgtRgr.SGD.tfw
0.0 8.0 3.0 0 LgtRgr.SGD.tfw
2 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
0.1 8.0 3.0 0 LgtRgr.SGD.tfw
4.0 1.0 1 LgtRgr.SGD.tfw
8.0 2.0 2 LgtRgr.SGD.tfw
4 LgtRgr.SGD.tfw
16.0 4.0 10 LgtRgr.SGD.tfw
0.0 8.0 7.0 10 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
8.0 5.0 10 LgtRgr.SGD.tfw
0.5 8.0 4.0 10 LgtRgr.SGD.tfw
16.0 4.0 10 LgtRgr.SGD.tfw
0.0 8.0 1.0 0 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
2 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
0.1 8.0 1.0 0 LgtRgr.SGD.tfw
0.0 32.0 1.0 10 LgtRgr.SGD.tfw
0.1 8.0 2.0 5 LgtRgr.SGD.tfw
0.0 8.0 0.5 10 LgtRgr.SGD.tfw
... ...
0.1 4.0 3.0 1 LgtRgr.SGD.tfw
16.0 3.0 10 LgtRgr.SGD.tfw
0.0 16.0 2.0 2 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
0 LgtRgr.SGD.tfw
100.0 18077.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
1000.0 18077.0 0.5 4.0 3.0 10 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 LgtRgr.tfw
0.1 0 LgtRgr.tfw
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
1000.0 9000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
8000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
7000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
6000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 LgtRgr.skl
1000.0 0.0 1000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 LgtRgr.tfw
1.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
100.0 1000.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 LgtRgr.tfw
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
logLossVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0.018463
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.193149
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 20.546733
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 19.209913
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 17.898932
2.0 1 20.091985
0.0 8.0 3.0 0 20.948740
2 20.948740
1 20.948740
10 20.948740
0.1 8.0 3.0 0 20.948740
4.0 1.0 1 19.602959
8.0 2.0 2 19.997834
4 20.635511
16.0 4.0 10 21.347350
0.0 8.0 7.0 10 21.886713
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 3.116799
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 22.423282
1 22.423282
8.0 5.0 10 22.167039
0.5 8.0 4.0 10 22.007029
16.0 4.0 10 21.561740
0.0 8.0 1.0 0 19.594075
1 19.594075
2 19.594075
10 19.594075
0.1 8.0 1.0 0 19.594075
0.0 32.0 1.0 10 16.835064
0.1 8.0 2.0 5 21.457194
0.0 8.0 0.5 10 16.048818
... ...
0.1 4.0 3.0 1 26.212732
16.0 3.0 10 25.966046
0.0 16.0 2.0 2 25.605293
10 25.605293
1 25.605293
0 25.605293
100.0 18077.0 0.1 16.0 1.0 10 24.085776
1000.0 18077.0 0.5 4.0 3.0 10 26.843734
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 3.085191
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 18.995890
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 20.546733
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 2.885114
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 9.502959
0.1 0 8.912652
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 24.085776
1000.0 9000.0 0.1 8.0 3.0 2 20.546733
8000.0 0.1 8.0 3.0 2 20.345963
7000.0 0.1 8.0 3.0 2 20.307083
6000.0 0.1 8.0 3.0 2 21.354761
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 2.610439
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 21.899301
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 2.498749
1000.0 0.0 1000.0 -1.0 0 2.539650
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 24.274683
1.0 0 14.714048
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 25.728109
100.0 1000.0 0.1 16.0 1.0 10 18.141110
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 12.600104
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 2.497940
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 25.837055
logLossVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'logLossCls': [0.0123335882818, 0.0241534716...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'logLossCls': [1.72596805748, 1.7853875361, ...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'logLossCls': [1.7477126908, 1.45845704706, ...
2.0 1 {u'logLossCls': [2.4098162987, 2.1249522721, 2...
0.0 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
2 {u'logLossCls': [1.99287215213, 1.25414623142,...
1 {u'logLossCls': [1.99287215213, 1.25414623142,...
10 {u'logLossCls': [1.99287215213, 1.25414623142,...
0.1 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
4.0 1.0 1 {u'logLossCls': [1.47759900586, 1.5216033003, ...
8.0 2.0 2 {u'logLossCls': [2.96027056113, 1.55809628297,...
4 {u'logLossCls': [2.1684078246, 2.40346788464, ...
16.0 4.0 10 {u'logLossCls': [1.79621628564, 1.99717570497,...
0.0 8.0 7.0 10 {u'logLossCls': [2.16842694926, 2.08737047825,...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'logLossCls': [1.82392002827, 1.3175497433, ...
1 {u'logLossCls': [1.82392002827, 1.3175497433, ...
8.0 5.0 10 {u'logLossCls': [2.49658612049, 3.01094383144,...
0.5 8.0 4.0 10 {u'logLossCls': [2.1293109356, 1.52307431523, ...
16.0 4.0 10 {u'logLossCls': [2.13819531033, 3.35150390363,...
0.0 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
1 {u'logLossCls': [2.37109926898, 2.31826985925,...
2 {u'logLossCls': [2.37109926898, 2.31826985925,...
10 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.1 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.0 32.0 1.0 10 {u'logLossCls': [1.3367673536, 2.19949472069, ...
0.1 8.0 2.0 5 {u'logLossCls': [2.47060812006, 2.55482495458,...
0.0 8.0 0.5 10 {u'logLossCls': [1.73155548863, 2.32786043541,...
... ...
0.1 4.0 3.0 1 {u'logLossCls': [1.71708309944, 3.4968377777, ...
16.0 3.0 10 {u'logLossCls': [1.71310413171, 3.4687167196, ...
0.0 16.0 2.0 2 {u'logLossCls': [1.86638576386, 3.42782085718,...
10 {u'logLossCls': [1.86638576386, 3.42782085718,...
1 {u'logLossCls': [1.86638576386, 3.42782085718,...
0 {u'logLossCls': [1.86638576386, 3.42782085718,...
100.0 18077.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 18077.0 0.5 4.0 3.0 10 {u'logLossCls': [2.62356084168, 3.42588099594,...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'logLossCls': [1.84540699409, 1.93896956479,...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'logLossCls': [1.14059053163, 1.46723165909,...
0.1 0 {u'logLossCls': [1.11536194976, 1.25718941188,...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 9000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
8000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.5257602538, 1.31972597838, ...
7000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.06754908637, 1.51088086827,...
6000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.97387028973, 1.20522205183,...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.57533202071, 1.59023941247,...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'logLossCls': [1.28781316316, 3.51519487621,...
1.0 0 {u'logLossCls': [0.842354858019, 1.17531354279...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.84808429073, 3.58756550329,...
100.0 1000.0 0.1 16.0 1.0 10 {u'logLossCls': [2.1826821181, 3.34540505739, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'logLossCls': [1.47411393905, 2.08387227918,...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'logLossCls': [3.67759601248, 3.22267129616,...
lrnRateTfw \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 10.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 3.0
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 10.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 1.0
2.0 1 2.0
0.0 8.0 3.0 0 3.0
2 3.0
1 3.0
10 3.0
0.1 8.0 3.0 0 3.0
4.0 1.0 1 1.0
8.0 2.0 2 2.0
4 2.0
16.0 4.0 10 4.0
0.0 8.0 7.0 10 7.0
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 4.0
1 4.0
8.0 5.0 10 5.0
0.5 8.0 4.0 10 4.0
16.0 4.0 10 4.0
0.0 8.0 1.0 0 1.0
1 1.0
2 1.0
10 1.0
0.1 8.0 1.0 0 1.0
0.0 32.0 1.0 10 1.0
0.1 8.0 2.0 5 2.0
0.0 8.0 0.5 10 0.5
... ...
0.1 4.0 3.0 1 3.0
16.0 3.0 10 3.0
0.0 16.0 2.0 2 2.0
10 2.0
1 2.0
0 2.0
100.0 18077.0 0.1 16.0 1.0 10 1.0
1000.0 18077.0 0.5 4.0 3.0 10 3.0
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 10.0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 3.0
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 1.0
0.1 0 0.1
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 1.0
1000.0 9000.0 0.1 8.0 3.0 2 3.0
8000.0 0.1 8.0 3.0 2 3.0
7000.0 0.1 8.0 3.0 2 3.0
6000.0 0.1 8.0 3.0 2 3.0
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 3.0
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 -1.0
1000.0 0.0 1000.0 -1.0 0 -1.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 10.0
1.0 0 1.0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 3.0
100.0 1000.0 0.1 16.0 1.0 10 1.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0.1
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 3.0
nObsBtc \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 22424.0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 22424.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 8.0
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 18077.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 8.0
2.0 1 8.0
0.0 8.0 3.0 0 8.0
2 8.0
1 8.0
10 8.0
0.1 8.0 3.0 0 8.0
4.0 1.0 1 4.0
8.0 2.0 2 8.0
4 8.0
16.0 4.0 10 16.0
0.0 8.0 7.0 10 8.0
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 18077.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 4.0
1 4.0
8.0 5.0 10 8.0
0.5 8.0 4.0 10 8.0
16.0 4.0 10 16.0
0.0 8.0 1.0 0 8.0
1 8.0
2 8.0
10 8.0
0.1 8.0 1.0 0 8.0
0.0 32.0 1.0 10 32.0
0.1 8.0 2.0 5 8.0
0.0 8.0 0.5 10 8.0
... ...
0.1 4.0 3.0 1 4.0
16.0 3.0 10 16.0
0.0 16.0 2.0 2 16.0
10 16.0
1 16.0
0 16.0
100.0 18077.0 0.1 16.0 1.0 10 16.0
1000.0 18077.0 0.5 4.0 3.0 10 4.0
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 15000.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 10000.0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 8.0
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 10000.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 10000.0
0.1 0 10000.0
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 16.0
1000.0 9000.0 0.1 8.0 3.0 2 8.0
8000.0 0.1 8.0 3.0 2 8.0
7000.0 0.1 8.0 3.0 2 8.0
6000.0 0.1 8.0 3.0 2 8.0
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 5000.0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 8.0
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 2000.0
1000.0 0.0 1000.0 -1.0 0 1000.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 1000.0
1.0 0 1000.0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 8.0
100.0 1000.0 0.1 16.0 1.0 10 16.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 1000.0
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 100.0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 8.0
nObsFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 22424.0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 22424.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 18077.0
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 18077.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 18077.0
2.0 1 18077.0
0.0 8.0 3.0 0 18077.0
2 18077.0
1 18077.0
10 18077.0
0.1 8.0 3.0 0 18077.0
4.0 1.0 1 18077.0
8.0 2.0 2 18077.0
4 18077.0
16.0 4.0 10 18077.0
0.0 8.0 7.0 10 18077.0
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 18077.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 18077.0
1 18077.0
8.0 5.0 10 18077.0
0.5 8.0 4.0 10 18077.0
16.0 4.0 10 18077.0
0.0 8.0 1.0 0 18077.0
1 18077.0
2 18077.0
10 18077.0
0.1 8.0 1.0 0 18077.0
0.0 32.0 1.0 10 18077.0
0.1 8.0 2.0 5 18077.0
0.0 8.0 0.5 10 18077.0
... ...
0.1 4.0 3.0 1 18077.0
16.0 3.0 10 18077.0
0.0 16.0 2.0 2 18077.0
10 18077.0
1 18077.0
0 18077.0
100.0 18077.0 0.1 16.0 1.0 10 18077.0
1000.0 18077.0 0.5 4.0 3.0 10 18077.0
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 15000.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 10000.0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 10000.0
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 10000.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 10000.0
0.1 0 10000.0
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 10000.0
1000.0 9000.0 0.1 8.0 3.0 2 9000.0
8000.0 0.1 8.0 3.0 2 8000.0
7000.0 0.1 8.0 3.0 2 7000.0
6000.0 0.1 8.0 3.0 2 6000.0
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 5000.0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 5000.0
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 2000.0
1000.0 0.0 1000.0 -1.0 0 1000.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 1000.0
1.0 0 1000.0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 1000.0
100.0 1000.0 0.1 16.0 1.0 10 1000.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 1000.0
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 100.0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 100.0
nStepsTfw \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 1000.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 1000.0
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 1000.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 1000.0
2.0 1 1000.0
0.0 8.0 3.0 0 1000.0
2 1000.0
1 1000.0
10 1000.0
0.1 8.0 3.0 0 1000.0
4.0 1.0 1 1000.0
8.0 2.0 2 1000.0
4 1000.0
16.0 4.0 10 1000.0
0.0 8.0 7.0 10 1000.0
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 1000.0
1 1000.0
8.0 5.0 10 1000.0
0.5 8.0 4.0 10 1000.0
16.0 4.0 10 1000.0
0.0 8.0 1.0 0 1000.0
1 1000.0
2 1000.0
10 1000.0
0.1 8.0 1.0 0 1000.0
0.0 32.0 1.0 10 1000.0
0.1 8.0 2.0 5 1000.0
0.0 8.0 0.5 10 1000.0
... ...
0.1 4.0 3.0 1 1000.0
16.0 3.0 10 1000.0
0.0 16.0 2.0 2 1000.0
10 1000.0
1 1000.0
0 1000.0
100.0 18077.0 0.1 16.0 1.0 10 100.0
1000.0 18077.0 0.5 4.0 3.0 10 1000.0
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 1000.0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 1000.0
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 -1.0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 1000.0
0.1 0 1000.0
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 100.0
1000.0 9000.0 0.1 8.0 3.0 2 1000.0
8000.0 0.1 8.0 3.0 2 1000.0
7000.0 0.1 8.0 3.0 2 1000.0
6000.0 0.1 8.0 3.0 2 1000.0
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 1000.0
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 -1.0
1000.0 0.0 1000.0 -1.0 0 -1.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 100.0
1.0 0 100.0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 1000.0
100.0 1000.0 0.1 16.0 1.0 10 100.0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 100.0
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 -1.0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 1000.0
predNew \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 NaN
2.0 1 NaN
0.0 8.0 3.0 0 NaN
2 NaN
1 NaN
10 NaN
0.1 8.0 3.0 0 NaN
4.0 1.0 1 NaN
8.0 2.0 2 NaN
4 NaN
16.0 4.0 10 NaN
0.0 8.0 7.0 10 NaN
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 NaN
1 NaN
8.0 5.0 10 NaN
0.5 8.0 4.0 10 NaN
16.0 4.0 10 NaN
0.0 8.0 1.0 0 NaN
1 NaN
2 NaN
10 NaN
0.1 8.0 1.0 0 NaN
0.0 32.0 1.0 10 NaN
0.1 8.0 2.0 5 NaN
0.0 8.0 0.5 10 NaN
... ...
0.1 4.0 3.0 1 NaN
16.0 3.0 10 NaN
0.0 16.0 2.0 2 NaN
10 NaN
1 NaN
0 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.5 4.0 3.0 10 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 NaN
0.1 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
1000.0 0.0 1000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 NaN
1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 NaN
100.0 1000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 NaN
rotateMaxAgl \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 2
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 2
2.0 1 1
0.0 8.0 3.0 0 0
2 2
1 1
10 10
0.1 8.0 3.0 0 0
4.0 1.0 1 1
8.0 2.0 2 2
4 4
16.0 4.0 10 10
0.0 8.0 7.0 10 10
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 10
1 1
8.0 5.0 10 10
0.5 8.0 4.0 10 10
16.0 4.0 10 10
0.0 8.0 1.0 0 0
1 1
2 2
10 10
0.1 8.0 1.0 0 0
0.0 32.0 1.0 10 10
0.1 8.0 2.0 5 5
0.0 8.0 0.5 10 10
... ...
0.1 4.0 3.0 1 1
16.0 3.0 10 10
0.0 16.0 2.0 2 2
10 10
1 1
0 0
100.0 18077.0 0.1 16.0 1.0 10 10
1000.0 18077.0 0.5 4.0 3.0 10 10
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 2
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 0
0.1 0 0
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 10
1000.0 9000.0 0.1 8.0 3.0 2 2
8000.0 0.1 8.0 3.0 2 2
7000.0 0.1 8.0 3.0 2 2
6000.0 0.1 8.0 3.0 2 2
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 2
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 0
1000.0 0.0 1000.0 -1.0 0 0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 0
1.0 0 0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 2
100.0 1000.0 0.1 16.0 1.0 10 10
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 2
rotatePby
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.1
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 0.1
2.0 1 0.1
0.0 8.0 3.0 0 0
2 0
1 0
10 0
0.1 8.0 3.0 0 0.1
4.0 1.0 1 0.1
8.0 2.0 2 0.1
4 0.1
16.0 4.0 10 0.1
0.0 8.0 7.0 10 0
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 0
1 0
8.0 5.0 10 0
0.5 8.0 4.0 10 0.5
16.0 4.0 10 0.5
0.0 8.0 1.0 0 0
1 0
2 0
10 0
0.1 8.0 1.0 0 0.1
0.0 32.0 1.0 10 0
0.1 8.0 2.0 5 0.1
0.0 8.0 0.5 10 0
... ...
0.1 4.0 3.0 1 0.1
16.0 3.0 10 0.1
0.0 16.0 2.0 2 0
10 0
1 0
0 0
100.0 18077.0 0.1 16.0 1.0 10 0.1
1000.0 18077.0 0.5 4.0 3.0 10 0.5
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 0.1
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 0
0.1 0 0
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 0.1
1000.0 9000.0 0.1 8.0 3.0 2 0.1
8000.0 0.1 8.0 3.0 2 0.1
7000.0 0.1 8.0 3.0 2 0.1
6000.0 0.1 8.0 3.0 2 0.1
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 0.1
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 0
1000.0 0.0 1000.0 -1.0 0 0
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 0
1.0 0 0
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 0.1
100.0 1000.0 0.1 16.0 1.0 10 0.1
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 0.1
[120 rows x 14 columns]
In [11]:
%run img_utils.py
srchParamsDct = {
'nObsFit' : [100, 1000, 5000, 6000, 7000, 8000, 9000, 10000, glbObsFitFtr.shape[0]],
# 'nObsFit' : [100, 1000, 5000, 10000,
# glbObsFitFtr.shape[0], glbObsTrnFtr.shape[0]]
'nObsBtc' : [8],
# 'nObsBtc' : [4, 8, 16, 32],
'rotatePby' : [0.1],
# 'rotatePby' : [0.0, 0.1, 0.2, 0.5],
'rotateMaxAgl' : [2],
# 'rotateMaxAgl' : [0, 1, 10],
'nStepsTfw' : [1000],
# 'nStepsTfw' : [100, 1000, 10000],
'lrnRateTfw' : [3.0]
# 'lrnRateTfw' : [0.1, 0.5, 1.0, 5.0, 7.0, 10.0]
}
jnk = mysearchParams(fitMdlLgtRgrSGDTfw, srchParamsDct = srchParamsDct,
curResultsDf = glbMdlDf,
mode = 'displayonly',
sort_values = ['nObsFit', 'accVld', 'logLossVld', 'duration'],
sort_ascending = [False , True , False, False],
save_drop_cols = 'model',
save_filepathname = glbPickleFile['models'],
lclXFit = glbXFit, lclYFit = glbYFit)
# thsDf, thsObsVldRspPredProba, thsObsNewRspPredProba = fitMdlLgtRgrTfw(
# glbXFit, glbYFit,
# nObsFit = 100, nStepsTfw = 10, lrnRateTfw = 0.5,
# visualize = False, newObs = False, verbose = False)
mysearchParams: will run <function fitMdlLgtRgrSGDTfw at 0x121d9bf50> with params: Empty DataFrame Columns: [nStepsTfw, nObsFit, rotatePby, nObsBtc, lrnRateTfw, rotateMaxAgl] Index: [] mysearchParams: total runs: 0
In [25]:
%run img_utils.py
glbMdlDf = mysearchParams(fitMdlLgtRgrSGDTfw, srchParamsDct = srchParamsDct,
curResultsDf = glbMdlDf,
mode = 'run',
sort_values = ['nObsFit', 'accVld', 'logLossVld', 'duration'],
sort_ascending = [False , False , True, True],
save_filepathname = glbPickleFile['models'],
save_drop_cols = 'model',
lclXFit = glbXFit, lclYFit = glbYFit)
/usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:124: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:143: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:126: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future
mysearchParams: running <function fitMdlLgtRgrSGDTfw at 0x1181d3f50> with params:
nStepsTfw 1000
nObsFit 6000
rotatePby 0.1
nObsBtc 8
lrnRateTfw 3
rotateMaxAgl 2
Logistic Regression (TensorFlow): nObsFit: 6000; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 1000; lrnRateTfw:3.0000
visualize: False; newObs: False; verbose: False
duration: 09 seconds
mysearchParams: running <function fitMdlLgtRgrSGDTfw at 0x1181d3f50> with params:
nStepsTfw 1000
nObsFit 7000
rotatePby 0.1
nObsBtc 8
lrnRateTfw 3
rotateMaxAgl 2
Logistic Regression (TensorFlow): nObsFit: 7000; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 1000; lrnRateTfw:3.0000
visualize: False; newObs: False; verbose: False
duration: 09 seconds
mysearchParams: running <function fitMdlLgtRgrSGDTfw at 0x1181d3f50> with params:
nStepsTfw 1000
nObsFit 8000
rotatePby 0.1
nObsBtc 8
lrnRateTfw 3
rotateMaxAgl 2
Logistic Regression (TensorFlow): nObsFit: 8000; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 1000; lrnRateTfw:3.0000
visualize: False; newObs: False; verbose: False
duration: 09 seconds
mysearchParams: running <function fitMdlLgtRgrSGDTfw at 0x1181d3f50> with params:
nStepsTfw 1000
nObsFit 9000
rotatePby 0.1
nObsBtc 8
lrnRateTfw 3
rotateMaxAgl 2
Logistic Regression (TensorFlow): nObsFit: 9000; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 1000; lrnRateTfw:3.0000
visualize: False; newObs: False; verbose: False
duration: 09 seconds
accVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'accCls': [0.989669421488, 0.991631799163, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'accCls': [0.504132231405, 0.435146443515, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'accCls': [0.448347107438, 0.44769874477, 0....
2.0 1 {u'accCls': [0.338842975207, 0.347280334728, 0...
0.0 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
2 {u'accCls': [0.460743801653, 0.640167364017, 0...
1 {u'accCls': [0.460743801653, 0.640167364017, 0...
10 {u'accCls': [0.460743801653, 0.640167364017, 0...
0.1 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
4.0 1.0 1 {u'accCls': [0.52479338843, 0.575313807531, 0....
8.0 2.0 2 {u'accCls': [0.146694214876, 0.5, 0.2123893805...
4 {u'accCls': [0.384297520661, 0.297071129707, 0...
16.0 4.0 10 {u'accCls': [0.510330578512, 0.389121338912, 0...
0.0 8.0 7.0 10 {u'accCls': [0.421487603306, 0.410041841004, 0...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'accCls': [0.506198347107, 0.644351464435, 0...
1 {u'accCls': [0.506198347107, 0.644351464435, 0...
8.0 5.0 10 {u'accCls': [0.338842975207, 0.173640167364, 0...
0.5 8.0 4.0 10 {u'accCls': [0.421487603306, 0.558577405858, 0...
16.0 4.0 10 {u'accCls': [0.396694214876, 0.0836820083682, ...
0.0 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
1 {u'accCls': [0.25, 0.246861924686, 0.303097345...
2 {u'accCls': [0.25, 0.246861924686, 0.303097345...
10 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.1 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.0 32.0 1.0 10 {u'accCls': [0.588842975207, 0.133891213389, 0...
0.1 8.0 2.0 5 {u'accCls': [0.276859504132, 0.219665271967, 0...
0.0 8.0 0.5 10 {u'accCls': [0.365702479339, 0.119246861925, 0...
... ...
0.1 4.0 3.0 1 {u'accCls': [0.545454545455, 0.0690376569038, ...
16.0 3.0 10 {u'accCls': [0.528925619835, 0.0523012552301, ...
0.0 16.0 2.0 2 {u'accCls': [0.495867768595, 0.0397489539749, ...
10 {u'accCls': [0.495867768595, 0.0397489539749, ...
1 {u'accCls': [0.495867768595, 0.0397489539749, ...
0 {u'accCls': [0.495867768595, 0.0397489539749, ...
100.0 18077.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 18077.0 0.5 4.0 3.0 10 {u'accCls': [0.297520661157, 0.081589958159, 0...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'accCls': [0.456611570248, 0.380753138075, 0...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'accCls': [0.289256198347, 0.182008368201, 0...
0.1 0 {u'accCls': [0.169421487603, 0.144351464435, 0...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 9000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
8000.0 0.1 8.0 3.0 2 {u'accCls': [0.553719008264, 0.610878661088, 0...
7000.0 0.1 8.0 3.0 2 {u'accCls': [0.431818181818, 0.558577405858, 0...
6000.0 0.1 8.0 3.0 2 {u'accCls': [0.181818181818, 0.671548117155, 0...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'accCls': [0.543388429752, 0.535564853556, 0...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'accCls': [0.646694214876, 0.0564853556485, ...
1.0 0 {u'accCls': [0.423553719008, 0.182008368201, 0...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'accCls': [0.204545454545, 0.0334728033473, ...
100.0 1000.0 0.1 16.0 1.0 10 {u'accCls': [0.136363636364, 0.0230125523013, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'accCls': [0.150826446281, 0.142259414226, 0...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'accCls': [0.0185950413223, 0.133891213389, ...
bestFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 False
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 True
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 False
2.0 1 False
0.0 8.0 3.0 0 False
2 False
1 False
10 False
0.1 8.0 3.0 0 False
4.0 1.0 1 False
8.0 2.0 2 False
4 False
16.0 4.0 10 False
0.0 8.0 7.0 10 False
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 False
1 False
8.0 5.0 10 False
0.5 8.0 4.0 10 False
16.0 4.0 10 False
0.0 8.0 1.0 0 False
1 False
2 False
10 False
0.1 8.0 1.0 0 False
0.0 32.0 1.0 10 False
0.1 8.0 2.0 5 False
0.0 8.0 0.5 10 False
... ...
0.1 4.0 3.0 1 False
16.0 3.0 10 False
0.0 16.0 2.0 2 False
10 False
1 False
0 False
100.0 18077.0 0.1 16.0 1.0 10 False
1000.0 18077.0 0.5 4.0 3.0 10 False
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 False
0.1 0 False
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 False
1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 False
1000.0 0.0 1000.0 -1.0 0 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 False
1.0 0 False
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 False
100.0 1000.0 0.1 16.0 1.0 10 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 False
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 False
logLossVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0.018463
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.193149
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 20.546733
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 19.209913
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 17.898932
2.0 1 20.091985
0.0 8.0 3.0 0 20.948740
2 20.948740
1 20.948740
10 20.948740
0.1 8.0 3.0 0 20.948740
4.0 1.0 1 19.602959
8.0 2.0 2 19.997834
4 20.635511
16.0 4.0 10 21.347350
0.0 8.0 7.0 10 21.886713
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 3.116799
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 22.423282
1 22.423282
8.0 5.0 10 22.167039
0.5 8.0 4.0 10 22.007029
16.0 4.0 10 21.561740
0.0 8.0 1.0 0 19.594075
1 19.594075
2 19.594075
10 19.594075
0.1 8.0 1.0 0 19.594075
0.0 32.0 1.0 10 16.835064
0.1 8.0 2.0 5 21.457194
0.0 8.0 0.5 10 16.048818
... ...
0.1 4.0 3.0 1 26.212732
16.0 3.0 10 25.966046
0.0 16.0 2.0 2 25.605293
10 25.605293
1 25.605293
0 25.605293
100.0 18077.0 0.1 16.0 1.0 10 24.085776
1000.0 18077.0 0.5 4.0 3.0 10 26.843734
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 3.085191
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 18.995890
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 20.546733
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 2.885114
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 9.502959
0.1 0 8.912652
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 24.085776
1000.0 9000.0 0.1 8.0 3.0 2 20.546733
8000.0 0.1 8.0 3.0 2 20.345963
7000.0 0.1 8.0 3.0 2 20.307083
6000.0 0.1 8.0 3.0 2 21.354761
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 2.610439
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 21.899301
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 2.498749
1000.0 0.0 1000.0 -1.0 0 2.539650
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 24.274683
1.0 0 14.714048
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 25.728109
100.0 1000.0 0.1 16.0 1.0 10 18.141110
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 12.600104
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 2.497940
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 25.837055
predNew \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 NaN
2.0 1 NaN
0.0 8.0 3.0 0 NaN
2 NaN
1 NaN
10 NaN
0.1 8.0 3.0 0 NaN
4.0 1.0 1 NaN
8.0 2.0 2 NaN
4 NaN
16.0 4.0 10 NaN
0.0 8.0 7.0 10 NaN
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 NaN
1 NaN
8.0 5.0 10 NaN
0.5 8.0 4.0 10 NaN
16.0 4.0 10 NaN
0.0 8.0 1.0 0 NaN
1 NaN
2 NaN
10 NaN
0.1 8.0 1.0 0 NaN
0.0 32.0 1.0 10 NaN
0.1 8.0 2.0 5 NaN
0.0 8.0 0.5 10 NaN
... ...
0.1 4.0 3.0 1 NaN
16.0 3.0 10 NaN
0.0 16.0 2.0 2 NaN
10 NaN
1 NaN
0 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.5 4.0 3.0 10 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 NaN
0.1 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
1000.0 0.0 1000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 NaN
1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 NaN
100.0 1000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 NaN
logLossVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'logLossCls': [0.0123335882818, 0.0241534716...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'logLossCls': [1.72596805748, 1.7853875361, ...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'logLossCls': [1.7477126908, 1.45845704706, ...
2.0 1 {u'logLossCls': [2.4098162987, 2.1249522721, 2...
0.0 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
2 {u'logLossCls': [1.99287215213, 1.25414623142,...
1 {u'logLossCls': [1.99287215213, 1.25414623142,...
10 {u'logLossCls': [1.99287215213, 1.25414623142,...
0.1 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
4.0 1.0 1 {u'logLossCls': [1.47759900586, 1.5216033003, ...
8.0 2.0 2 {u'logLossCls': [2.96027056113, 1.55809628297,...
4 {u'logLossCls': [2.1684078246, 2.40346788464, ...
16.0 4.0 10 {u'logLossCls': [1.79621628564, 1.99717570497,...
0.0 8.0 7.0 10 {u'logLossCls': [2.16842694926, 2.08737047825,...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'logLossCls': [1.82392002827, 1.3175497433, ...
1 {u'logLossCls': [1.82392002827, 1.3175497433, ...
8.0 5.0 10 {u'logLossCls': [2.49658612049, 3.01094383144,...
0.5 8.0 4.0 10 {u'logLossCls': [2.1293109356, 1.52307431523, ...
16.0 4.0 10 {u'logLossCls': [2.13819531033, 3.35150390363,...
0.0 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
1 {u'logLossCls': [2.37109926898, 2.31826985925,...
2 {u'logLossCls': [2.37109926898, 2.31826985925,...
10 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.1 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.0 32.0 1.0 10 {u'logLossCls': [1.3367673536, 2.19949472069, ...
0.1 8.0 2.0 5 {u'logLossCls': [2.47060812006, 2.55482495458,...
0.0 8.0 0.5 10 {u'logLossCls': [1.73155548863, 2.32786043541,...
... ...
0.1 4.0 3.0 1 {u'logLossCls': [1.71708309944, 3.4968377777, ...
16.0 3.0 10 {u'logLossCls': [1.71310413171, 3.4687167196, ...
0.0 16.0 2.0 2 {u'logLossCls': [1.86638576386, 3.42782085718,...
10 {u'logLossCls': [1.86638576386, 3.42782085718,...
1 {u'logLossCls': [1.86638576386, 3.42782085718,...
0 {u'logLossCls': [1.86638576386, 3.42782085718,...
100.0 18077.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 18077.0 0.5 4.0 3.0 10 {u'logLossCls': [2.62356084168, 3.42588099594,...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'logLossCls': [1.84540699409, 1.93896956479,...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'logLossCls': [1.14059053163, 1.46723165909,...
0.1 0 {u'logLossCls': [1.11536194976, 1.25718941188,...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 9000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
8000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.5257602538, 1.31972597838, ...
7000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.06754908637, 1.51088086827,...
6000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.97387028973, 1.20522205183,...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.57533202071, 1.59023941247,...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'logLossCls': [1.28781316316, 3.51519487621,...
1.0 0 {u'logLossCls': [0.842354858019, 1.17531354279...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.84808429073, 3.58756550329,...
100.0 1000.0 0.1 16.0 1.0 10 {u'logLossCls': [2.1826821181, 3.34540505739, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'logLossCls': [1.47411393905, 2.08387227918,...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'logLossCls': [3.67759601248, 3.22267129616,...
accVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 1.000000
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.987348
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.381643
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0.375201
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 0.361859
2.0 1 0.357488
0.0 8.0 3.0 0 0.357028
2 0.357028
1 0.357028
10 0.357028
0.1 8.0 3.0 0 0.357028
4.0 1.0 1 0.355648
8.0 2.0 2 0.351737
4 0.348516
16.0 4.0 10 0.346216
0.0 8.0 7.0 10 0.346216
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 0.343685
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 0.338854
1 0.338854
8.0 5.0 10 0.336784
0.5 8.0 4.0 10 0.334484
16.0 4.0 10 0.331723
0.0 8.0 1.0 0 0.331493
1 0.331493
2 0.331493
10 0.331493
0.1 8.0 1.0 0 0.331493
0.0 32.0 1.0 10 0.328042
0.1 8.0 2.0 5 0.325512
0.0 8.0 0.5 10 0.325052
... ...
0.1 4.0 3.0 1 0.226823
16.0 3.0 10 0.219692
0.0 16.0 2.0 2 0.214861
10 0.214861
1 0.214861
0 0.214861
100.0 18077.0 0.1 16.0 1.0 10 0.210950
1000.0 18077.0 0.5 4.0 3.0 10 0.208420
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0.342535
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0.384173
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 0.381643
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0.358638
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 0.319991
0.1 0 0.251208
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 0.210950
1000.0 9000.0 0.1 8.0 3.0 2 0.381643
8000.0 0.1 8.0 3.0 2 0.378882
7000.0 0.1 8.0 3.0 2 0.375431
6000.0 0.1 8.0 3.0 2 0.339545
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0.363699
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 0.320911
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 0.341385
1000.0 0.0 1000.0 -1.0 0 0.333333
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 0.273522
1.0 0 0.255809
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 0.224983
100.0 1000.0 0.1 16.0 1.0 10 0.212790
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0.173453
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0.307108
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 0.224753
duration \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 874
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 743
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 11
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 519
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 9
2.0 1 10
0.0 8.0 3.0 0 8
2 8
1 9
10 10
0.1 8.0 3.0 0 10
4.0 1.0 1 8
8.0 2.0 2 10
4 13
16.0 4.0 10 14
0.0 8.0 7.0 10 9
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 597
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 8
1 8
8.0 5.0 10 9
0.5 8.0 4.0 10 15
16.0 4.0 10 24
0.0 8.0 1.0 0 8
1 8
2 8
10 9
0.1 8.0 1.0 0 13
0.0 32.0 1.0 10 16
0.1 8.0 2.0 5 10
0.0 8.0 0.5 10 9
... ...
0.1 4.0 3.0 1 9
16.0 3.0 10 13
0.0 16.0 2.0 2 10
10 11
1 11
0 11
100.0 18077.0 0.1 16.0 1.0 10 2
1000.0 18077.0 0.5 4.0 3.0 10 12
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 514
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 288
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 317
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 301
0.1 0 295
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 2
1000.0 9000.0 0.1 8.0 3.0 2 9
8000.0 0.1 8.0 3.0 2 9
7000.0 0.1 8.0 3.0 2 9
6000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 150
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 55
1000.0 0.0 1000.0 -1.0 0 28
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 7
1.0 0 5
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 9
100.0 1000.0 0.1 16.0 1.0 10 2
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 5
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 8
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 9
model
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 NaN
2.0 1 NaN
0.0 8.0 3.0 0 NaN
2 NaN
1 NaN
10 NaN
0.1 8.0 3.0 0 NaN
4.0 1.0 1 NaN
8.0 2.0 2 NaN
4 NaN
16.0 4.0 10 NaN
0.0 8.0 7.0 10 NaN
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 NaN
1 NaN
8.0 5.0 10 NaN
0.5 8.0 4.0 10 NaN
16.0 4.0 10 NaN
0.0 8.0 1.0 0 NaN
1 NaN
2 NaN
10 NaN
0.1 8.0 1.0 0 NaN
0.0 32.0 1.0 10 NaN
0.1 8.0 2.0 5 NaN
0.0 8.0 0.5 10 NaN
... ...
0.1 4.0 3.0 1 NaN
16.0 3.0 10 NaN
0.0 16.0 2.0 2 NaN
10 NaN
1 NaN
0 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.5 4.0 3.0 10 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 NaN
0.1 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
1000.0 9000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
8000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
7000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
6000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 NaN
1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
100.0 1000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
[120 rows x 8 columns]
Compressed pickle file: data/img_M_SFDD_ImgSz_64.pickle; size: 41 KB
In [16]:
lrn10Srs = glbMdlDf.ix[('LgtRgr.tfw', 1000.0, 18077.0, 10.0)]
# print lrn10Srs
lrn01Srs = glbMdlDf.ix[('LgtRgr.tfw', 1000.0, 18077.0, 1.0)]
# print lrn01Srs
print "lrn10Srs: accVld: %.4f" % (lrn10Srs['accVld'])
print "lrn01Srs: accVld: %.4f" % (lrn01Srs['accVld'])
print "lrn10-01: accVldDff: %.4f" % (lrn10Srs['accVld'] - lrn01Srs['accVld'])
print "lrn10Srs: accVldCls:"; print (lrn10Srs['accVldCls']['accCls'])
print "lrn01Srs: accVldCls:"; print (lrn01Srs['accVldCls']['accCls'])
print "lrn10-01: accVldClsDff:"; print (lrn10Srs['accVldCls']['accCls'] -
lrn01Srs['accVldCls']['accCls'])
print "\n"
print "lrn10Srs: logLossVld: %8.4f" % (lrn10Srs['logLossVld'])
print "lrn01Srs: logLossVld: %8.4f" % (lrn01Srs['logLossVld'])
print "lrn10-01: logLossVldDff: %8.4f" % (lrn10Srs['logLossVld'] -
lrn01Srs['logLossVld'])
print "lrn10Srs: logLossVldCls:"; print (lrn10Srs['logLossVldCls']['logLossCls'])
print "lrn01Srs: logLossVldCls:"; print (lrn01Srs['logLossVldCls']['logLossCls'])
print "lrn10-01: logLossVldClsDff:";
print (lrn10Srs['logLossVldCls']['logLossCls'] -
lrn01Srs['logLossVldCls']['logLossCls'])
lrn10Srs: accVld: 0.3752 lrn01Srs: accVld: 0.3202 lrn10-01: accVldDff: 0.0550 lrn10Srs: accVldCls: [ 0.5041 0.4351 0.3805 0.1738 0.3965 0.2464 0.6749 0.5143 0.25 0.1637] lrn01Srs: accVldCls: [ 0.3264 0.2092 0.2323 0.2144 0.3878 0.1588 0.6637 0.6286 0.2088 0.219 ] lrn10-01: accVldClsDff: [ 0.1777 0.2259 0.1482 -0.0406 0.0087 0.0877 0.0113 -0.1143 0.0412 -0.0553] lrn10Srs: logLossVld: 19.2099 lrn01Srs: logLossVld: 9.4750 lrn10-01: logLossVldDff: 9.7349 lrn10Srs: logLossVldCls: [ 1.726 1.7854 1.8043 2.6157 1.9122 2.3841 0.9763 1.0939 1.9704 2.9416] lrn01Srs: logLossVldCls: [ 1.0565 1.4437 0.9226 1.0162 0.8034 1.5807 0.2146 0.2664 0.879 1.2919] lrn10-01: logLossVldClsDff: [ 0.6695 0.3417 0.8817 1.5995 1.1088 0.8034 0.7616 0.8275 1.0915 1.6497]
In [43]:
# Set value based on condition
# print glbMdlDf.ix[glbMdlDf['id'].str.contains('LogisticRegression.SGD.tf',
# na=False), 'id']
# glbMdlDf.ix[glbMdlDf['id'].str.contains('LogisticRegression.SGD.tf',
# na=False), 'id'] = 'LgtRgr.SGD.tf'
# print glbMdlDf.ix[glbMdlDf['id'].str.contains('LogisticRegression.SGD.tf',
# na=False), 'id']
def lclfixNanDf(df, column, default):
print "\n Before:"
dspDf = df[[column]]
dspDf[column + '.isnull'] = df[column].isnull()
print dspDf.to_string(index = False)
df.loc[df[column].isnull(), column] = default
print "\n After:"
dspDf = df[[column]]
dspDf[column + '.isnull'] = df[column].isnull()
print dspDf.to_string(index = False)
return(df)
# tmpMdlDf = lclfixNanDf(glbMdlDf, column = 'rotatePby', default = 0)
# print glbMdlDf.ix[- glbMdlDf['nStepsTfw'].isnull(), ['id', 'nStepsTfw']]
# glbMdlDf.ix[- glbMdlDf['nStepsTfw'].isnull(), 'id'] = 'LogisticRegression.tf'
# print glbMdlDf.ix[- glbMdlDf['nStepsTfw'].isnull(), ['id', 'nStepsTfw']]
# print glbMdlDf.ix[glbMdlDf['nObsBtc'].isnull(), ['nObsBtc', 'nObsFit']]
# # glbMdlDf.ix[glbMdlDf['nObsBtc'].isnull(), 'nObsBtc'] = \
# # glbMdlDf.ix[glbMdlDf['nObsBtc'].isnull(), 'nObsFit']
# glbMdlDf['nObsBtc'] = glbMdlDf.apply(
# lambda (row): row['nObsFit'] if pd.isnull(row['nObsBtc']) else row['nObsBtc'],
# axis = 1)
# print 'After:'
# print glbMdlDf.ix[glbMdlDf['nObsBtc'].isnull(), ['nObsBtc', 'nObsFit']]
# Change value
# tmpMdlDf = glbMdlDf
# print tmpMdlDf[(tmpMdlDf['id'].str.contains('LgtRgr.skl', na = False)) &
# (tmpMdlDf['nStepsTfw'] == 1.0)]
# print tmpMdlDf.ix[(tmpMdlDf['id'].str.contains('LgtRgr.skl', na = False)) &
# (tmpMdlDf['nStepsTfw'] == 1.0), 'nStepsTfw']
# tmpMdlDf.ix[(tmpMdlDf['id'].str.contains('LgtRgr.skl', na = False)) &
# (tmpMdlDf['nStepsTfw'] == 1.0), 'nStepsTfw'] = -1.0
# print 'After:'
# print tmpMdlDf
# Remove specific models
# mask = (glbMdlDf['id'].str.contains('LgtRgr.tfw', na = False))
# mask = (glbMdlDf['id'].str.contains('LgtRgr.tfw', na = False)) & \
# (glbMdlDf['nObsFit'] == 10000.0)
# print mask
# tmpMdlDf = glbMdlDf[~mask]
# print tmpMdlDf
# Remove dups
# print glbMdlDf.columns
# print (glbMdlDf['logLossVld'])
# print (glbMdlDf.index.duplicated())
# tmpMdlDf = glbMdlDf[~glbMdlDf.index.duplicated()]
# print (tmpMdlDf.index.duplicated())
# glbMdlDf['nObsBtc'] = glbMdlDf['nObsFit']
# glbMdlDf = tmpMdlDf
# print 'After:'
# print glbMdlDf
print glbMdlDf['bestFit']
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 False
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 NaN
18077.0 0.0 18077.0 10.0 0 True
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 False
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 1.0 0 False
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 False
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 False
0.1 0 False
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 False
2000.0 0.0 2000.0 -1.0 0 False
1000.0 0.0 1000.0 -1.0 0 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 False
1.0 0 False
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 False
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 False
Name: bestFit, dtype: object
In [ ]:
In [37]:
myexportDf(glbMdlDf,
save_filepathname = glbPickleFile['models'],
save_drop_cols = 'model'
)
Compressed pickle file: data/img_M_SFDD_ImgSz_64.pickle; size: 7 KB
In [12]:
glbMdlDf['bestFit'] = False
glbMdlDf.ix[(
'LgtRgr.SGD.tfw', 1000.0, 18077.0, 0.1, 8.0, 3.0, 2),
# LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2
# id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
'bestFit'] = True
print glbMdlDf[list(set(glbMdlDf.columns) - set(srchParamsDct.keys()))]
# print glbMdlDf[glbMdlDf.nObsFit >= 10000][
# list(set(glbMdlDf.columns) - set(srchParamsDct.keys()))]
accVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'accCls': [0.989669421488, 0.991631799163, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'accCls': [0.504132231405, 0.435146443515, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'accCls': [0.448347107438, 0.44769874477, 0....
2.0 1 {u'accCls': [0.338842975207, 0.347280334728, 0...
0.0 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
2 {u'accCls': [0.460743801653, 0.640167364017, 0...
1 {u'accCls': [0.460743801653, 0.640167364017, 0...
10 {u'accCls': [0.460743801653, 0.640167364017, 0...
0.1 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
4.0 1.0 1 {u'accCls': [0.52479338843, 0.575313807531, 0....
8.0 2.0 2 {u'accCls': [0.146694214876, 0.5, 0.2123893805...
4 {u'accCls': [0.384297520661, 0.297071129707, 0...
16.0 4.0 10 {u'accCls': [0.510330578512, 0.389121338912, 0...
0.0 8.0 7.0 10 {u'accCls': [0.421487603306, 0.410041841004, 0...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'accCls': [0.506198347107, 0.644351464435, 0...
1 {u'accCls': [0.506198347107, 0.644351464435, 0...
8.0 5.0 10 {u'accCls': [0.338842975207, 0.173640167364, 0...
0.5 8.0 4.0 10 {u'accCls': [0.421487603306, 0.558577405858, 0...
16.0 4.0 10 {u'accCls': [0.396694214876, 0.0836820083682, ...
0.0 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
1 {u'accCls': [0.25, 0.246861924686, 0.303097345...
2 {u'accCls': [0.25, 0.246861924686, 0.303097345...
10 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.1 8.0 1.0 0 {u'accCls': [0.25, 0.246861924686, 0.303097345...
0.0 32.0 1.0 10 {u'accCls': [0.588842975207, 0.133891213389, 0...
0.1 8.0 2.0 5 {u'accCls': [0.276859504132, 0.219665271967, 0...
0.0 8.0 0.5 10 {u'accCls': [0.365702479339, 0.119246861925, 0...
... ...
0.1 4.0 3.0 1 {u'accCls': [0.545454545455, 0.0690376569038, ...
16.0 3.0 10 {u'accCls': [0.528925619835, 0.0523012552301, ...
0.0 16.0 2.0 2 {u'accCls': [0.495867768595, 0.0397489539749, ...
10 {u'accCls': [0.495867768595, 0.0397489539749, ...
1 {u'accCls': [0.495867768595, 0.0397489539749, ...
0 {u'accCls': [0.495867768595, 0.0397489539749, ...
100.0 18077.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 18077.0 0.5 4.0 3.0 10 {u'accCls': [0.297520661157, 0.081589958159, 0...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'accCls': [0.456611570248, 0.380753138075, 0...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'accCls': [0.289256198347, 0.182008368201, 0...
0.1 0 {u'accCls': [0.169421487603, 0.144351464435, 0...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 9000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
8000.0 0.1 8.0 3.0 2 {u'accCls': [0.553719008264, 0.610878661088, 0...
7000.0 0.1 8.0 3.0 2 {u'accCls': [0.431818181818, 0.558577405858, 0...
6000.0 0.1 8.0 3.0 2 {u'accCls': [0.181818181818, 0.671548117155, 0...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'accCls': [0.543388429752, 0.535564853556, 0...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'accCls': [0.646694214876, 0.0564853556485, ...
1.0 0 {u'accCls': [0.423553719008, 0.182008368201, 0...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'accCls': [0.204545454545, 0.0334728033473, ...
100.0 1000.0 0.1 16.0 1.0 10 {u'accCls': [0.136363636364, 0.0230125523013, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'accCls': [0.150826446281, 0.142259414226, 0...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'accCls': [0.0185950413223, 0.133891213389, ...
bestFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 False
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 True
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 False
2.0 1 False
0.0 8.0 3.0 0 False
2 False
1 False
10 False
0.1 8.0 3.0 0 False
4.0 1.0 1 False
8.0 2.0 2 False
4 False
16.0 4.0 10 False
0.0 8.0 7.0 10 False
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 False
1 False
8.0 5.0 10 False
0.5 8.0 4.0 10 False
16.0 4.0 10 False
0.0 8.0 1.0 0 False
1 False
2 False
10 False
0.1 8.0 1.0 0 False
0.0 32.0 1.0 10 False
0.1 8.0 2.0 5 False
0.0 8.0 0.5 10 False
... ...
0.1 4.0 3.0 1 False
16.0 3.0 10 False
0.0 16.0 2.0 2 False
10 False
1 False
0 False
100.0 18077.0 0.1 16.0 1.0 10 False
1000.0 18077.0 0.5 4.0 3.0 10 False
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 False
0.1 0 False
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 False
1000.0 9000.0 0.1 8.0 3.0 2 False
8000.0 0.1 8.0 3.0 2 False
7000.0 0.1 8.0 3.0 2 False
6000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 False
1000.0 0.0 1000.0 -1.0 0 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 False
1.0 0 False
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 False
100.0 1000.0 0.1 16.0 1.0 10 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 False
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 False
logLossVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0.018463
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.193149
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 20.546733
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 19.209913
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 17.898932
2.0 1 20.091985
0.0 8.0 3.0 0 20.948740
2 20.948740
1 20.948740
10 20.948740
0.1 8.0 3.0 0 20.948740
4.0 1.0 1 19.602959
8.0 2.0 2 19.997834
4 20.635511
16.0 4.0 10 21.347350
0.0 8.0 7.0 10 21.886713
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 3.116799
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 22.423282
1 22.423282
8.0 5.0 10 22.167039
0.5 8.0 4.0 10 22.007029
16.0 4.0 10 21.561740
0.0 8.0 1.0 0 19.594075
1 19.594075
2 19.594075
10 19.594075
0.1 8.0 1.0 0 19.594075
0.0 32.0 1.0 10 16.835064
0.1 8.0 2.0 5 21.457194
0.0 8.0 0.5 10 16.048818
... ...
0.1 4.0 3.0 1 26.212732
16.0 3.0 10 25.966046
0.0 16.0 2.0 2 25.605293
10 25.605293
1 25.605293
0 25.605293
100.0 18077.0 0.1 16.0 1.0 10 24.085776
1000.0 18077.0 0.5 4.0 3.0 10 26.843734
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 3.085191
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 18.995890
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 20.546733
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 2.885114
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 9.502959
0.1 0 8.912652
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 24.085776
1000.0 9000.0 0.1 8.0 3.0 2 20.546733
8000.0 0.1 8.0 3.0 2 20.345963
7000.0 0.1 8.0 3.0 2 20.307083
6000.0 0.1 8.0 3.0 2 21.354761
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 2.610439
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 21.899301
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 2.498749
1000.0 0.0 1000.0 -1.0 0 2.539650
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 24.274683
1.0 0 14.714048
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 25.728109
100.0 1000.0 0.1 16.0 1.0 10 18.141110
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 12.600104
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 2.497940
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 25.837055
predNew \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 NaN
2.0 1 NaN
0.0 8.0 3.0 0 NaN
2 NaN
1 NaN
10 NaN
0.1 8.0 3.0 0 NaN
4.0 1.0 1 NaN
8.0 2.0 2 NaN
4 NaN
16.0 4.0 10 NaN
0.0 8.0 7.0 10 NaN
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 NaN
1 NaN
8.0 5.0 10 NaN
0.5 8.0 4.0 10 NaN
16.0 4.0 10 NaN
0.0 8.0 1.0 0 NaN
1 NaN
2 NaN
10 NaN
0.1 8.0 1.0 0 NaN
0.0 32.0 1.0 10 NaN
0.1 8.0 2.0 5 NaN
0.0 8.0 0.5 10 NaN
... ...
0.1 4.0 3.0 1 NaN
16.0 3.0 10 NaN
0.0 16.0 2.0 2 NaN
10 NaN
1 NaN
0 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.5 4.0 3.0 10 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 NaN
0.1 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
1000.0 0.0 1000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 NaN
1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 NaN
100.0 1000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 NaN
logLossVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'logLossCls': [0.0123335882818, 0.0241534716...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'logLossCls': [1.72596805748, 1.7853875361, ...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 {u'logLossCls': [1.7477126908, 1.45845704706, ...
2.0 1 {u'logLossCls': [2.4098162987, 2.1249522721, 2...
0.0 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
2 {u'logLossCls': [1.99287215213, 1.25414623142,...
1 {u'logLossCls': [1.99287215213, 1.25414623142,...
10 {u'logLossCls': [1.99287215213, 1.25414623142,...
0.1 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
4.0 1.0 1 {u'logLossCls': [1.47759900586, 1.5216033003, ...
8.0 2.0 2 {u'logLossCls': [2.96027056113, 1.55809628297,...
4 {u'logLossCls': [2.1684078246, 2.40346788464, ...
16.0 4.0 10 {u'logLossCls': [1.79621628564, 1.99717570497,...
0.0 8.0 7.0 10 {u'logLossCls': [2.16842694926, 2.08737047825,...
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 {u'logLossCls': [1.82392002827, 1.3175497433, ...
1 {u'logLossCls': [1.82392002827, 1.3175497433, ...
8.0 5.0 10 {u'logLossCls': [2.49658612049, 3.01094383144,...
0.5 8.0 4.0 10 {u'logLossCls': [2.1293109356, 1.52307431523, ...
16.0 4.0 10 {u'logLossCls': [2.13819531033, 3.35150390363,...
0.0 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
1 {u'logLossCls': [2.37109926898, 2.31826985925,...
2 {u'logLossCls': [2.37109926898, 2.31826985925,...
10 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.1 8.0 1.0 0 {u'logLossCls': [2.37109926898, 2.31826985925,...
0.0 32.0 1.0 10 {u'logLossCls': [1.3367673536, 2.19949472069, ...
0.1 8.0 2.0 5 {u'logLossCls': [2.47060812006, 2.55482495458,...
0.0 8.0 0.5 10 {u'logLossCls': [1.73155548863, 2.32786043541,...
... ...
0.1 4.0 3.0 1 {u'logLossCls': [1.71708309944, 3.4968377777, ...
16.0 3.0 10 {u'logLossCls': [1.71310413171, 3.4687167196, ...
0.0 16.0 2.0 2 {u'logLossCls': [1.86638576386, 3.42782085718,...
10 {u'logLossCls': [1.86638576386, 3.42782085718,...
1 {u'logLossCls': [1.86638576386, 3.42782085718,...
0 {u'logLossCls': [1.86638576386, 3.42782085718,...
100.0 18077.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 18077.0 0.5 4.0 3.0 10 {u'logLossCls': [2.62356084168, 3.42588099594,...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'logLossCls': [1.84540699409, 1.93896956479,...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 {u'logLossCls': [1.14059053163, 1.46723165909,...
0.1 0 {u'logLossCls': [1.11536194976, 1.25718941188,...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 9000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
8000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.5257602538, 1.31972597838, ...
7000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.06754908637, 1.51088086827,...
6000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.97387028973, 1.20522205183,...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.57533202071, 1.59023941247,...
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 NaN
1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 {u'logLossCls': [1.28781316316, 3.51519487621,...
1.0 0 {u'logLossCls': [0.842354858019, 1.17531354279...
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.84808429073, 3.58756550329,...
100.0 1000.0 0.1 16.0 1.0 10 {u'logLossCls': [2.1826821181, 3.34540505739, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'logLossCls': [1.47411393905, 2.08387227918,...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'logLossCls': [3.67759601248, 3.22267129616,...
accVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 1.000000
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.987348
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.381643
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0.375201
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 0.361859
2.0 1 0.357488
0.0 8.0 3.0 0 0.357028
2 0.357028
1 0.357028
10 0.357028
0.1 8.0 3.0 0 0.357028
4.0 1.0 1 0.355648
8.0 2.0 2 0.351737
4 0.348516
16.0 4.0 10 0.346216
0.0 8.0 7.0 10 0.346216
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 0.343685
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 0.338854
1 0.338854
8.0 5.0 10 0.336784
0.5 8.0 4.0 10 0.334484
16.0 4.0 10 0.331723
0.0 8.0 1.0 0 0.331493
1 0.331493
2 0.331493
10 0.331493
0.1 8.0 1.0 0 0.331493
0.0 32.0 1.0 10 0.328042
0.1 8.0 2.0 5 0.325512
0.0 8.0 0.5 10 0.325052
... ...
0.1 4.0 3.0 1 0.226823
16.0 3.0 10 0.219692
0.0 16.0 2.0 2 0.214861
10 0.214861
1 0.214861
0 0.214861
100.0 18077.0 0.1 16.0 1.0 10 0.210950
1000.0 18077.0 0.5 4.0 3.0 10 0.208420
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0.342535
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0.384173
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 0.381643
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0.358638
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 0.319991
0.1 0 0.251208
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 0.210950
1000.0 9000.0 0.1 8.0 3.0 2 0.381643
8000.0 0.1 8.0 3.0 2 0.378882
7000.0 0.1 8.0 3.0 2 0.375431
6000.0 0.1 8.0 3.0 2 0.339545
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0.363699
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 0.320911
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 0.341385
1000.0 0.0 1000.0 -1.0 0 0.333333
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 0.273522
1.0 0 0.255809
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 0.224983
100.0 1000.0 0.1 16.0 1.0 10 0.212790
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0.173453
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0.307108
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 0.224753
duration \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 874
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 743
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 11
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 519
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 9
2.0 1 10
0.0 8.0 3.0 0 8
2 8
1 9
10 10
0.1 8.0 3.0 0 10
4.0 1.0 1 8
8.0 2.0 2 10
4 13
16.0 4.0 10 14
0.0 8.0 7.0 10 9
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 597
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 8
1 8
8.0 5.0 10 9
0.5 8.0 4.0 10 15
16.0 4.0 10 24
0.0 8.0 1.0 0 8
1 8
2 8
10 9
0.1 8.0 1.0 0 13
0.0 32.0 1.0 10 16
0.1 8.0 2.0 5 10
0.0 8.0 0.5 10 9
... ...
0.1 4.0 3.0 1 9
16.0 3.0 10 13
0.0 16.0 2.0 2 10
10 11
1 11
0 11
100.0 18077.0 0.1 16.0 1.0 10 2
1000.0 18077.0 0.5 4.0 3.0 10 12
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 514
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 288
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 317
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 301
0.1 0 295
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 2
1000.0 9000.0 0.1 8.0 3.0 2 9
8000.0 0.1 8.0 3.0 2 9
7000.0 0.1 8.0 3.0 2 9
6000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 150
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 55
1000.0 0.0 1000.0 -1.0 0 28
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 7
1.0 0 5
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 9
100.0 1000.0 0.1 16.0 1.0 10 2
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 5
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 8
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 9
id
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 1.0 2 LgtRgr.SGD.tfw
2.0 1 LgtRgr.SGD.tfw
0.0 8.0 3.0 0 LgtRgr.SGD.tfw
2 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
0.1 8.0 3.0 0 LgtRgr.SGD.tfw
4.0 1.0 1 LgtRgr.SGD.tfw
8.0 2.0 2 LgtRgr.SGD.tfw
4 LgtRgr.SGD.tfw
16.0 4.0 10 LgtRgr.SGD.tfw
0.0 8.0 7.0 10 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 18077.0 0.0 18077.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 18077.0 0.0 4.0 4.0 10 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
8.0 5.0 10 LgtRgr.SGD.tfw
0.5 8.0 4.0 10 LgtRgr.SGD.tfw
16.0 4.0 10 LgtRgr.SGD.tfw
0.0 8.0 1.0 0 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
2 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
0.1 8.0 1.0 0 LgtRgr.SGD.tfw
0.0 32.0 1.0 10 LgtRgr.SGD.tfw
0.1 8.0 2.0 5 LgtRgr.SGD.tfw
0.0 8.0 0.5 10 LgtRgr.SGD.tfw
... ...
0.1 4.0 3.0 1 LgtRgr.SGD.tfw
16.0 3.0 10 LgtRgr.SGD.tfw
0.0 16.0 2.0 2 LgtRgr.SGD.tfw
10 LgtRgr.SGD.tfw
1 LgtRgr.SGD.tfw
0 LgtRgr.SGD.tfw
100.0 18077.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
1000.0 18077.0 0.5 4.0 3.0 10 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 1.0 0 LgtRgr.tfw
0.1 0 LgtRgr.tfw
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
1000.0 9000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
8000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
7000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
6000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 5000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
LgtRgr.skl -1.0 2000.0 0.0 2000.0 -1.0 0 LgtRgr.skl
1000.0 0.0 1000.0 -1.0 0 LgtRgr.skl
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 10.0 0 LgtRgr.tfw
1.0 0 LgtRgr.tfw
LgtRgr.SGD.tfw 1000.0 1000.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
100.0 1000.0 0.1 16.0 1.0 10 LgtRgr.SGD.tfw
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 LgtRgr.tfw
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 LgtRgr.skl
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
[120 rows x 8 columns]
/usr/local/lib/python2.7/site-packages/ipykernel/kernelbase.py:212: PerformanceWarning: indexing past lexsort depth may impact performance. handler(stream, idents, msg)
In [4]:
# # With gradient descent training, even this much data is prohibitive.
# # Subset the training data for faster turnaround.
# nObsFit = 10000
# graph = tf.Graph()
# with graph.as_default():
# # Input data.
# # Load the training, validation and test data into constants that are
# # attached to the graph.
# tfwObsFitFtr = tf.constant(glbXFit[:nObsFit, :])
# tfwObsFitRsp = tf.constant(glbYFit[:nObsFit])
# tfwObsVldFtr = tf.constant(glbXVld)
# tfwObsNewFtr = tf.constant(glbXNew)
# # Variables.
# # These are the parameters that we are going to be training. The weight
# # matrix will be initialized using random valued following a (truncated)
# # normal distribution. The tfwB get initialized to zero.
# tfwW = tf.Variable(
# tf.truncated_normal([glbImg['size'] * glbImg['size'], glbRspClassN]), name = 'tfwW')
# tfwB = tf.Variable(tf.zeros([glbRspClassN]), name = 'tfwB')
# print(tfwW.initialized_value())
# print(tfwB.initialized_value())
# # print 'initial tfwB:%s' % (np.vectorize("%.4e".__mod__)(tf.get_variable('tfwB')))
# # print 'initial tfwW (first 5 only):'
# # for lblIx in xrange(glbRspClassN):
# # print 'lblIx:%2d:%s'% (np.vectorize("%.4e".__mod__)(tfwW.value()[:5, lblIx]))
# # Training computation.
# # We multiply the inputs with the weight matrix, and add tfwB. We compute
# # the softmax and cross-entropy (it's one operation in TensorFlow, because
# # it's very common, and it can be optimized). We take the average of this
# # cross-entropy across all training examples: that's our loss.
# logits = tf.matmul(tfwObsFitFtr, tfwW) + tfwB
# loss = tf.reduce_mean(
# tf.nn.softmax_cross_entropy_with_logits(logits, tfwObsFitRsp))
# # Optimizer.
# # We are going to find the minimum of this loss using gradient descent.
# optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# # Predictions for the training, validation, and test data.
# # These are not part of training, but merely here so that we can report
# # accuracy_score figures as we train.
# tfwObsTrnPred = tf.nn.softmax(logits)
# tfwObsVldPred = tf.nn.softmax(
# tf.matmul(tfwObsVldFtr, tfwW) + tfwB)
# tfwObsNewPred = tf.nn.softmax(tf.matmul(tfwObsNewFtr, tfwW) + tfwB)
Tensor("Identity:0", shape=TensorShape([Dimension(784), Dimension(10)]), dtype=float32)
Tensor("Identity_1:0", shape=TensorShape([Dimension(10)]), dtype=float32)
Let's run this computation and iterate:
In [58]:
# nStepsTfw = 801
# def accuracy_score(predictions, labels):
# return (100.0 * np.sum(np.argmax(predictions, 1) == np.argmax(labels, 1))
# / predictions.shape[0])
# with tf.Session(graph=graph) as session:
# # This is a one-time operation which ensures the parameters get initialized as
# # we described in the graph: random tfwW for the matrix, zeros for the
# # tfwB.
# tf.initialize_all_variables().run()
# print('Initialized')
# for step in range(nStepsTfw):
# # Run the computations. We tell .run() that we want to run the optimizer,
# # and get the loss value and the training predictions returned as numpy
# # arrays.
# _, l, predictions = session.run([optimizer, loss, tfwObsTrnPred])
# if (step % 100 == 0):
# print('Loss at step %d: %f' % (step, l))
# print('Training accuracy_score: %.1f%%' % accuracy_score(
# predictions, glbYFit[:nObsFit, :]))
# # Calling .eval() on tfwObsVldPred is basically like calling run(), but
# # just to get that one numpy array. Note that it recomputes all its graph
# # dependencies.
# print('Validation accuracy_score: %.1f%%' % accuracy_score(
# tfwObsVldPred.eval(), glbYVld))
# print('Test accuracy_score: %.1f%%' % accuracy_score(tfwObsNewPred.eval(), glbYNew))
In [13]:
robjects.pandas2ri.activate()
pltRDf = robjects.conversion.py2ri(glbMdlDf)
# print(pltRDf)
pltRFn = robjects.r("""
source('~/Dropbox/datascience/R/myplot.R')
function(RDf, filename) {
mypltModelStats(RDf, c('accVld', 'logLossVld', 'duration'),
dim = c('nObsFit', 'id', 'nStepsTfw', 'lrnRateTfw'),
scaleXFn = NULL,
#highLightIx = which.min(RDf$logLossVld),
highLightIx = which(RDf$bestFit == 'TRUE'),
title = NULL,
fileName = filename)
}
""")
pltRFn(pltRDf, 'img_04_fit_lgtRgr_SGD_Tfw_SFDD_glbMdlDf.png')
pltRFn = robjects.r("""
source('~/Dropbox/datascience/R/myplot.R')
function(RDf, filename) {
mypltModelStats(RDf, c('accVld', 'logLossVld'),
dim = c('nObsFit', 'id', 'nStepsTfw', 'lrnRateTfw'),
scaleXFn = NULL,
#highLightIx = which.min(RDf$logLossVld),
highLightIx = which(RDf$bestFit == 'TRUE'),
title = NULL,
fileName = filename)
}
""")
pltRFn(pltRDf, 'img_04_fit_lgtRgr_SGD_Tfw_SFDD_glbMdlDf_logLossVld.png')
pltRFn = robjects.r("""
source('~/Dropbox/datascience/R/myplot.R')
function(RDf, filename) {
mypltModelStats(RDf, c('accVld'),
dim = c('nObsFit', 'id',
'nObsBtc', 'rotatePby', 'rotateMaxAgl',
'nStepsTfw', 'lrnRateTfw'),
scaleXFn = NULL,
#highLightIx = which.min(RDf$logLossVld),
highLightIx = which(RDf$bestFit == 'TRUE'),
title = NULL,
fileName = filename)
}
""")
pltRFn(pltRDf, 'img_04_fit_lgtRgr_SGD_Tfw_SFDD_glbMdlDf_accVld.png')
pltRFn = robjects.r("""
source('~/Dropbox/datascience/R/myplot.R')
function(RDf, filename) {
mypltModelStats(subset(RDf,
# nObsBtc %in% c(8, 18077, 22424) &
lrnRateTfw %in% c(-1.0, 3.0, 10.0) &
rotatePby %in% c(0.0, 0.1) &
rotateMaxAgl %in% c(0, 2)
),
measure = c('accVld'),
dim = c('nObsFit', 'id',
# 'nObsBtc',
# 'lrnRateTfw',
# 'rotatePby',
'rotateMaxAgl',
'nStepsTfw',
NULL),
scaleXFn = NULL,
#highLightIx = which.min(RDf$logLossVld),
highLightIx = which(RDf$bestFit == 'TRUE'),
title = NULL,
fileName = filename)
}
""")
pltRFn(pltRDf, 'img_04_fit_lgtRgr_SGD_Tfw_SFDD_glbMdlDf_accVldSel.png')
# id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
/usr/local/lib/python2.7/site-packages/rpy2/robjects/functions.py:106: UserWarning: geom_path: Each group consists of only one observation. Do you need to adjust the group aesthetic? res = super(Function, self).__call__(*new_args, **new_kwargs)
Out[13]:
<ListVector - Python:0x13c88f320 / R:0x7fe0225c8bf8> [DataF..., ListV..., Envir..., ..., ListV..., Envir..., ListV...] <ListVector - Python:0x13c88f320 / R:0x7fe0225c8bf8> [DataF..., ListV..., Envir..., ..., ListV..., Envir..., ListV...] <ListVector - Python:0x13c88f320 / R:0x7fe0225c8bf8> [DataF..., ListV..., Envir..., ..., ListV..., Envir..., ListV...] scales: <class 'rpy2.robjects.environments.Environment'> <Environment - Python:0x13c869710 / R:0x7fe00b93f778> ... <ListVector - Python:0x13c88f320 / R:0x7fe0225c8bf8> [DataF..., ListV..., Envir..., ..., ListV..., Envir..., ListV...] layers: <class 'rpy2.robjects.environments.Environment'> <Environment - Python:0x13c88f098 / R:0x7fe00c39dc08> <ListVector - Python:0x13c88f320 / R:0x7fe0225c8bf8> [DataF..., ListV..., Envir..., ..., ListV..., Envir..., ListV...]
In [29]:
glbMdlDf.to_csv('img_04_fit_lgtRgr_SGD_Tfw_SFDD_glbMdlDf.csv')
Fit selected model to glbObsFit¶
In [14]:
selMdlSrs = glbMdlDf[glbMdlDf['bestFit']]
print selMdlSrs
accVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.381643
accVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
bestFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 True
duration \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 11
id \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 LgtRgr.SGD.tfw
logLossVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 20.546733
logLossVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
lrnRateTfw \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 3.0
nObsBtc \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 8.0
nObsFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 18077.0
nStepsTfw \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 1000.0
predNew \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
rotateMaxAgl \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 2
rotatePby
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.1
In [15]:
selMdlDf, selYVldPby, selYNewPby = fitMdlLgtRgrSGDTfw(
glbXFit, glbYFit,
nObsFit = glbXFit.shape[0],
nObsBtc = selMdlSrs['nObsBtc'][0],
rotatePby = selMdlSrs['rotatePby'][0],
rotateMaxAgl = selMdlSrs['rotateMaxAgl'][0],
nStepsTfw = selMdlSrs['nStepsTfw'][0],
lrnRateTfw = selMdlSrs['lrnRateTfw'][0],
visualize = True, newObs = True, verbose = True)
# def fitMdlLgtRgrSGDTfw(lclXFit, lclYFit,
# nObsFit = 100, nObsBtc = 16,
# rotatePby = 0.1, rotateMaxAgl = 5,
# nStepsTfw = 10, lrnRateTfw = 0.1,
# visualize = False, newObs = False, verbose = False):
Logistic Regression (TensorFlow): nObsFit:18077; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 1000; lrnRateTfw:3.0000
visualize: True; newObs: True; verbose: True
(' tfwW:', <tf.Tensor 'Identity:0' shape=(4096, 10) dtype=float32>)
(' tfwB:', <tf.Tensor 'Identity_1:0' shape=(10,) dtype=float32>)
Initialized
/usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:124: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:143: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:126: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/numpy/core/fromnumeric.py:225: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future return reshape(newshape, order=order)
step 1( 0 secs): Minibatch rotation:angle: 0.7815 step 6( 0 secs): Minibatch rotation:angle: 1.3900 step 13( 0 secs): Minibatch rotation:angle: 1.4487 step 34( 0 secs): Minibatch rotation:angle: 0.2647 step 39( 0 secs): Minibatch rotation:angle: -0.4238 step 40( 0 secs): Minibatch rotation:angle: -1.0228 step 42( 1 secs): Minibatch rotation:angle: -1.5997 step 73( 1 secs): Minibatch rotation:angle: -1.4536 step 75( 1 secs): Minibatch rotation:angle: 0.7455 step 78( 1 secs): Minibatch rotation:angle: -0.2561 step 86( 1 secs): Minibatch rotation:angle: 0.0559 step 89( 1 secs): Minibatch rotation:angle: 0.8329 step 93( 1 secs): Minibatch rotation:angle: 0.1879 step 101( 1 secs): Minibatch rotation:angle: 0.1547 step 123( 1 secs): Minibatch rotation:angle: 0.4447 step 134( 1 secs): Minibatch rotation:angle: -1.7246 step 139( 1 secs): Minibatch rotation:angle: -0.6370 step 148( 1 secs): Minibatch rotation:angle: -0.7290 step 149( 1 secs): Minibatch rotation:angle: 0.6298 step 168( 1 secs): Minibatch rotation:angle: 0.9559 step 174( 1 secs): Minibatch rotation:angle: 1.4031 step 177( 1 secs): Minibatch rotation:angle: 1.5177 step 193( 1 secs): Minibatch rotation:angle: -1.5264 step 208( 2 secs): Minibatch rotation:angle: 0.8651 step 219( 2 secs): Minibatch rotation:angle: -0.7559 step 227( 2 secs): Minibatch rotation:angle: -0.7687 step 228( 2 secs): Minibatch rotation:angle: 0.7853 step 229( 2 secs): Minibatch rotation:angle: 0.7060 step 259( 2 secs): Minibatch rotation:angle: 1.1189 step 262( 2 secs): Minibatch rotation:angle: 1.2409 step 284( 2 secs): Minibatch rotation:angle: 1.2055 step 302( 2 secs): Minibatch rotation:angle: -0.8014 step 307( 2 secs): Minibatch rotation:angle: 1.5852 step 320( 2 secs): Minibatch rotation:angle: -1.7845 step 323( 2 secs): Minibatch rotation:angle: 0.6667 step 355( 2 secs): Minibatch rotation:angle: -0.7972 step 356( 2 secs): Minibatch rotation:angle: -1.0431 step 380( 2 secs): Minibatch rotation:angle: 0.8782 step 401( 4 secs): Minibatch rotation:angle: 1.0272 step 406( 4 secs): Minibatch rotation:angle: 1.2680 step 412( 4 secs): Minibatch rotation:angle: -0.3036 step 427( 4 secs): Minibatch rotation:angle: 0.6773 step 456( 4 secs): Minibatch rotation:angle: -0.2512 step 457( 4 secs): Minibatch rotation:angle: 1.2892 step 462( 4 secs): Minibatch rotation:angle: -1.4782 step 520( 4 secs): Minibatch rotation:angle: 0.1854 step 524( 4 secs): Minibatch rotation:angle: -0.2769 step 544( 4 secs): Minibatch rotation:angle: -0.3829 step 552( 4 secs): Minibatch rotation:angle: -0.8067 step 555( 4 secs): Minibatch rotation:angle: -1.0371 step 565( 4 secs): Minibatch rotation:angle: -0.0422 step 581( 4 secs): Minibatch rotation:angle: -1.8645 step 583( 4 secs): Minibatch rotation:angle: -0.3721 step 587( 4 secs): Minibatch rotation:angle: -0.6783 step 633( 6 secs): Minibatch rotation:angle: 0.2982 step 644( 6 secs): Minibatch rotation:angle: -0.1618 step 657( 6 secs): Minibatch rotation:angle: -0.7546 step 658( 6 secs): Minibatch rotation:angle: 0.1138 step 681( 6 secs): Minibatch rotation:angle: -1.7727 step 682( 6 secs): Minibatch rotation:angle: 0.7245 step 687( 6 secs): Minibatch rotation:angle: 1.4412 step 689( 6 secs): Minibatch rotation:angle: -0.5182 step 691( 6 secs): Minibatch rotation:angle: 0.6104 step 700( 6 secs): Minibatch rotation:angle: 0.2777 step 713( 6 secs): Minibatch rotation:angle: 0.4088 step 731( 6 secs): Minibatch rotation:angle: 1.5995 step 761( 6 secs): Minibatch rotation:angle: -0.1049 step 764( 6 secs): Minibatch rotation:angle: -0.1286 step 785( 6 secs): Minibatch rotation:angle: 0.2296 step 795( 6 secs): Minibatch rotation:angle: -1.4570 step 813( 7 secs): Minibatch rotation:angle: -1.5104 step 837( 7 secs): Minibatch rotation:angle: 0.9490 step 847( 7 secs): Minibatch rotation:angle: 1.2389 step 857( 7 secs): Minibatch rotation:angle: 1.9708 step 861( 7 secs): Minibatch rotation:angle: -1.3822 step 877( 7 secs): Minibatch rotation:angle: -1.6069 step 880( 7 secs): Minibatch rotation:angle: 1.4276 step 882( 7 secs): Minibatch rotation:angle: 0.0323 step 895( 7 secs): Minibatch rotation:angle: -0.7840 step 909( 7 secs): Minibatch rotation:angle: 0.2200 step 924( 7 secs): Minibatch rotation:angle: 0.9365 step 939( 7 secs): Minibatch rotation:angle: 0.5616 step 942( 7 secs): Minibatch rotation:angle: -1.3131 step 945( 7 secs): Minibatch rotation:angle: 0.5496 step 948( 7 secs): Minibatch rotation:angle: 1.8078 step 952( 7 secs): Minibatch rotation:angle: 1.6791 step 954( 7 secs): Minibatch rotation:angle: 1.3377 step 969( 7 secs): Minibatch rotation:angle: 0.3555 step 981( 7 secs): Minibatch rotation:angle: -0.8890 step 992( 7 secs): Minibatch rotation:angle: -0.7087 Vld accuracy:0.3816 [ 0.6074 0.2197 0.5951 0.5643 0.1939 0.1896 0.456 0.6771 0.1841 0.146 ] [[294 9 75 9 64 0 14 19 0 0] [ 3 105 103 1 0 0 7 243 1 15] [ 0 1 269 11 0 0 99 65 7 0] [ 21 0 79 250 58 25 0 2 2 6] [ 63 1 83 157 89 19 3 9 13 22] [ 73 14 14 44 13 80 133 27 24 0] [ 2 9 109 72 1 0 202 36 12 0] [ 0 4 49 16 0 2 34 237 8 0] [ 14 2 97 18 3 5 68 89 67 1] [ 11 0 181 115 52 5 1 20 1 66]]
Vld logLoss:20.5467 [ 1.4886 2.7646 1.3374 1.4457 2.908 2.6567 1.7421 0.815 2.3394 3.0493] max Pby for cls: c0; desc: normal driving; proba: 1.0000; nObs: 421 img_62023.jpg: plot_occlusion:
display_weight:
Proba: [ 1. 0. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c0; desc: normal driving; proba: 0.6046; nObs: 1 img_3899.jpg: plot_occlusion:
display_weight:
Proba: [ 0.6046 0. 0. 0. 0.3954 0. 0. 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c1; desc: texting - right; proba: 1.0000; nObs: 115 img_53130.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c1; desc: texting - right; proba: 0.6158; nObs: 1 img_100358.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.6158 0.3842 0. 0. 0. 0. 0. 0. 0. ] next best class: talking on the phone - right max Pby for cls: c2; desc: talking on the phone - right; proba: 1.0000; nObs: 928 img_24934.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c2; desc: talking on the phone - right; proba: 0.5197; nObs: 1 img_52596.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.5197 0. 0. 0. 0.4803 0. 0. 0. ] next best class: drinking max Pby for cls: c3; desc: texting - left; proba: 1.0000; nObs: 621 img_79904.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.] min Pby for cls: c3; desc: texting - left; proba: 0.5347; nObs: 1 img_42027.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.5347 0. 0. 0.4653 0. 0. 0. ] next best class: drinking max Pby for cls: c4; desc: talking on the phone - left; proba: 1.0000; nObs: 261 img_18343.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.] min Pby for cls: c4; desc: talking on the phone - left; proba: 0.7097; nObs: 1 img_27020.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0.7097 0. 0. 0.2903 0. 0. ] next best class: reaching behind max Pby for cls: c5; desc: operating the radio; proba: 1.0000; nObs: 113 img_15928.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.] min Pby for cls: c5; desc: operating the radio; proba: 0.5496; nObs: 1 img_13512.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.4504 0. 0.5496 0. 0. 0. 0. ] next best class: texting - left max Pby for cls: c6; desc: drinking; proba: 1.0000; nObs: 480 img_29251.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.] min Pby for cls: c6; desc: drinking; proba: 0.5265; nObs: 1 img_43784.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0.4735 0.5265 0. 0. 0. ] next best class: operating the radio max Pby for cls: c7; desc: reaching behind; proba: 1.0000; nObs: 644 img_80068.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.] min Pby for cls: c7; desc: reaching behind; proba: 0.5231; nObs: 1 img_30223.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.4769 0. 0. 0. 0. 0.5231 0. 0. ] next best class: talking on the phone - right max Pby for cls: c8; desc: hair and makeup; proba: 1.0000; nObs: 114 img_31586.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.] min Pby for cls: c8; desc: hair and makeup; proba: 0.5134; nObs: 1 img_63085.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.4866 0. 0. 0. 0. 0. 0.5134 0. ] next best class: talking on the phone - right max Pby for cls: c9; desc: talking to passenger; proba: 1.0000; nObs: 91 img_25562.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.] min Pby for cls: c9; desc: talking to passenger; proba: 0.5606; nObs: 1 img_61126.jpg: plot_occlusion:
display_weight:
Proba: [ 0.4394 0. 0. 0. 0. 0. 0. 0. 0. 0.5606] next best class: normal driving predicting 79726 new obs... max Pby for cls: c0; desc: normal driving; proba: 1.0000; nObs: 9879 img_100.jpg: plot_occlusion:
display_weight:
Proba: [ 1. 0. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c0; desc: normal driving; proba: 0.4028; nObs: 1 img_19394.jpg: plot_occlusion:
display_weight:
Proba: [ 0.4028 0. 0. 0.367 0. 0. 0. 0. 0. 0.2302] next best class: texting - left max Pby for cls: c1; desc: texting - right; proba: 1.0000; nObs: 2864 img_100030.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c1; desc: texting - right; proba: 0.5020; nObs: 1 img_47748.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.502 0. 0.498 0. 0. 0. 0. 0. 0. ] next best class: texting - left max Pby for cls: c2; desc: talking on the phone - right; proba: 1.0000; nObs: 11332 img_1.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c2; desc: talking on the phone - right; proba: 0.5030; nObs: 1 img_8923.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.503 0. 0. 0. 0. 0.497 0. 0. ] next best class: reaching behind max Pby for cls: c3; desc: texting - left; proba: 1.0000; nObs: 5611 img_10.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.] min Pby for cls: c3; desc: texting - left; proba: 0.4147; nObs: 1 img_10687.jpg: plot_occlusion:
display_weight:
Proba: [ 0.2919 0. 0. 0.4147 0. 0. 0. 0.2934 0. 0. ] next best class: reaching behind max Pby for cls: c4; desc: talking on the phone - left; proba: 1.0000; nObs: 2413 img_100016.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.] min Pby for cls: c4; desc: talking on the phone - left; proba: 0.4235; nObs: 1 img_73467.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.2586 0.4235 0. 0. 0.3179 0. 0. ] next best class: reaching behind max Pby for cls: c5; desc: operating the radio; proba: 1.0000; nObs: 4406 img_100023.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.] min Pby for cls: c5; desc: operating the radio; proba: 0.5054; nObs: 1 img_38145.jpg: plot_occlusion:
display_weight:
Proba: [ 0.4946 0. 0. 0. 0. 0.5054 0. 0. 0. 0. ] next best class: normal driving max Pby for cls: c6; desc: drinking; proba: 1.0000; nObs: 5603 img_100004.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.] min Pby for cls: c6; desc: drinking; proba: 0.5026; nObs: 1 img_92615.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0.5026 0. 0.4974 0. ] next best class: hair and makeup max Pby for cls: c7; desc: reaching behind; proba: 1.0000; nObs: 12661 img_100005.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.] min Pby for cls: c7; desc: reaching behind; proba: 0.4289; nObs: 1 img_32173.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.2325 0. 0. 0. 0.3386 0.4289 0. 0. ] next best class: drinking max Pby for cls: c8; desc: hair and makeup; proba: 1.0000; nObs: 5467 img_100007.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.] min Pby for cls: c8; desc: hair and makeup; proba: 0.4734; nObs: 1 img_97640.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.1803 0. 0. 0. 0. 0.3463 0. 0.4734 0. ] next best class: drinking max Pby for cls: c9; desc: talking to passenger; proba: 1.0000; nObs: 6562 img_100028.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.] min Pby for cls: c9; desc: talking to passenger; proba: 0.4141; nObs: 1 img_9380.jpg: plot_occlusion:
display_weight:
Proba:
[ 0.2825 0.3034 0. 0. 0. 0. 0. 0. 0.
0.4141]
next best class: texting - right
New prediction knts:
{'kntCls': (array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]), array([11609, 3451, 13417, 6688, 3000, 4989, 6714, 14866, 6579, 8413]))}
duration: 107 seconds
In [16]:
print '\n selMdl:'
glbwriteSubmission(glbINew, selYNewPby,
'img_04_fit_lgtRgr_SGD_Tfw_SFDD_ImgSz_' + str(glbImg['size']) + \
'_sbmt_sel.csv')
selMdl:
img c0 c1 c2 c3 c4 \
img
img_1.jpg img_1.jpg 6.764812e-32 0.0 1.0 0.0 3.422848e-29
img_10.jpg img_10.jpg 0.000000e+00 0.0 0.0 1.0 0.000000e+00
img_100.jpg img_100.jpg 1.000000e+00 0.0 0.0 0.0 0.000000e+00
img_1000.jpg img_1000.jpg 2.215907e-07 0.0 0.0 0.0 8.074588e-30
img_100000.jpg img_100000.jpg 0.000000e+00 0.0 0.0 1.0 9.966320e-19
c5 c6 c7 c8 c9
img
img_1.jpg 0.000000e+00 0.000000e+00 3.077862e-11 0.0 0.0
img_10.jpg 8.291059e-29 0.000000e+00 0.000000e+00 0.0 0.0
img_100.jpg 0.000000e+00 0.000000e+00 0.000000e+00 0.0 0.0
img_1000.jpg 0.000000e+00 5.287736e-14 0.000000e+00 0.0 1.0
img_100000.jpg 0.000000e+00 0.000000e+00 0.000000e+00 0.0 0.0
img c0 c1 c2 c3 \
img
img_99994.jpg img_99994.jpg 0.980845 0.019155 0.000000e+00 0.000000e+00
img_99995.jpg img_99995.jpg 0.000000 0.000000 3.444694e-31 0.000000e+00
img_99996.jpg img_99996.jpg 0.000000 0.000000 4.330597e-16 1.721270e-19
img_99998.jpg img_99998.jpg 0.000000 0.000000 1.061564e-27 0.000000e+00
img_99999.jpg img_99999.jpg 0.000000 0.000000 0.000000e+00 0.000000e+00
c4 c5 c6 c7 c8 \
img
img_99994.jpg 0.0 3.311643e-38 0.000000e+00 1.641042e-21 0.0
img_99995.jpg 0.0 0.000000e+00 9.911789e-21 0.000000e+00 0.0
img_99996.jpg 1.0 0.000000e+00 0.000000e+00 0.000000e+00 0.0
img_99998.jpg 0.0 0.000000e+00 1.000000e+00 1.118691e-16 0.0
img_99999.jpg 0.0 0.000000e+00 1.000000e+00 0.000000e+00 0.0
c9
img
img_99994.jpg 5.080267e-27
img_99995.jpg 1.000000e+00
img_99996.jpg 0.000000e+00
img_99998.jpg 0.000000e+00
img_99999.jpg 0.000000e+00
exporting 79726 rows to img_04_fit_lgtRgr_SGD_Tfw_SFDD_ImgSz_64_sbmt_sel.csv...
Fit selected model to glbObsTrn¶
In [24]:
finMdlDf, finYVldPby, finYNewPby = fitMdlLgtRgrSGDTfw(
glbXTrn, glbYTrn,
nObsFit = glbXTrn.shape[0],
nObsBtc = selMdlSrs['nObsBtc'][0],
rotatePby = selMdlSrs['rotatePby'][0],
rotateMaxAgl = selMdlSrs['rotateMaxAgl'][0],
# Ensure all Trn obs are used at least once
nStepsTfw = max(selMdlSrs['nStepsTfw'][0],
glbXTrn.shape[0] * 1.0 / selMdlSrs['nObsBtc'][0] + 1),
lrnRateTfw = selMdlSrs['lrnRateTfw'][0],
visualize = True, newObs = True, verbose = True)
# selMdlDf, selYVldPby, selYNewPby = fitMdlLgtRgrSGDTfw(
# glbXFit, glbYFit,
# nObsFit = glbXFit.shape[0],
# nObsBtc = selMdlSrs['nObsBtc'][0],
# rotatePby = selMdlSrs['rotatePby'][0],
# rotateMaxAgl = selMdlSrs['rotateMaxAgl'][0],
# nStepsTfw = selMdlSrs['nStepsTfw'][0],
# lrnRateTfw = selMdlSrs['lrnRateTfw'][0],
# visualize = True, newObs = True, verbose = True)
Logistic Regression (TensorFlow): nObsFit:22424; nObsBtc: 8; rotatePby: 0.1000; rotateMaxAgl: 2; nStepsTfw: 2804; lrnRateTfw:3.0000
visualize: True; newObs: True; verbose: True
(' tfwW:', <tf.Tensor 'Identity:0' shape=(4096, 10) dtype=float32>)
(' tfwB:', <tf.Tensor 'Identity_1:0' shape=(10,) dtype=float32>)
Initialized
/usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:124: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:143: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future /usr/local/lib/python2.7/site-packages/ipykernel/__main__.py:126: VisibleDeprecationWarning: using a non-integer number instead of an integer will result in an error in the future
step 1( 0 secs): Minibatch rotation:angle: 0.7815 step 6( 0 secs): Minibatch rotation:angle: 1.3900 step 13( 0 secs): Minibatch rotation:angle: 1.4487 step 34( 0 secs): Minibatch rotation:angle: 0.2647 step 39( 0 secs): Minibatch rotation:angle: -0.4238 step 40( 0 secs): Minibatch rotation:angle: -1.0228 step 42( 0 secs): Minibatch rotation:angle: -1.5997 step 73( 1 secs): Minibatch rotation:angle: -1.4536 step 75( 1 secs): Minibatch rotation:angle: 0.7455 step 78( 1 secs): Minibatch rotation:angle: -0.2561 step 86( 1 secs): Minibatch rotation:angle: 0.0559 step 89( 1 secs): Minibatch rotation:angle: 0.8329 step 93( 1 secs): Minibatch rotation:angle: 0.1879 step 101( 1 secs): Minibatch rotation:angle: 0.1547 step 123( 1 secs): Minibatch rotation:angle: 0.4447 step 134( 1 secs): Minibatch rotation:angle: -1.7246 step 139( 1 secs): Minibatch rotation:angle: -0.6370 step 148( 1 secs): Minibatch rotation:angle: -0.7290 step 149( 1 secs): Minibatch rotation:angle: 0.6298 step 168( 1 secs): Minibatch rotation:angle: 0.9559 step 174( 1 secs): Minibatch rotation:angle: 1.4031 step 177( 1 secs): Minibatch rotation:angle: 1.5177 step 193( 1 secs): Minibatch rotation:angle: -1.5264 step 208( 2 secs): Minibatch rotation:angle: 0.8651 step 219( 2 secs): Minibatch rotation:angle: -0.7559 step 227( 2 secs): Minibatch rotation:angle: -0.7687 step 228( 2 secs): Minibatch rotation:angle: 0.7853 step 229( 2 secs): Minibatch rotation:angle: 0.7060 step 259( 2 secs): Minibatch rotation:angle: 1.1189 step 262( 2 secs): Minibatch rotation:angle: 1.2409 step 284( 2 secs): Minibatch rotation:angle: 1.2055 step 302( 2 secs): Minibatch rotation:angle: -0.8014 step 307( 2 secs): Minibatch rotation:angle: 1.5852 step 320( 2 secs): Minibatch rotation:angle: -1.7845 step 323( 2 secs): Minibatch rotation:angle: 0.6667 step 355( 2 secs): Minibatch rotation:angle: -0.7972 step 356( 2 secs): Minibatch rotation:angle: -1.0431 step 380( 2 secs): Minibatch rotation:angle: 0.8782 step 401( 4 secs): Minibatch rotation:angle: 1.0272 step 406( 4 secs): Minibatch rotation:angle: 1.2680 step 412( 4 secs): Minibatch rotation:angle: -0.3036 step 427( 4 secs): Minibatch rotation:angle: 0.6773 step 456( 4 secs): Minibatch rotation:angle: -0.2512 step 457( 4 secs): Minibatch rotation:angle: 1.2892 step 462( 4 secs): Minibatch rotation:angle: -1.4782 step 520( 4 secs): Minibatch rotation:angle: 0.1854 step 524( 4 secs): Minibatch rotation:angle: -0.2769 step 544( 4 secs): Minibatch rotation:angle: -0.3829 step 552( 4 secs): Minibatch rotation:angle: -0.8067 step 555( 4 secs): Minibatch rotation:angle: -1.0371 step 565( 4 secs): Minibatch rotation:angle: -0.0422 step 581( 4 secs): Minibatch rotation:angle: -1.8645 step 583( 4 secs): Minibatch rotation:angle: -0.3721 step 587( 4 secs): Minibatch rotation:angle: -0.6783 step 633( 5 secs): Minibatch rotation:angle: 0.2982 step 644( 5 secs): Minibatch rotation:angle: -0.1618 step 657( 5 secs): Minibatch rotation:angle: -0.7546 step 658( 5 secs): Minibatch rotation:angle: 0.1138 step 681( 5 secs): Minibatch rotation:angle: -1.7727 step 682( 5 secs): Minibatch rotation:angle: 0.7245 step 687( 5 secs): Minibatch rotation:angle: 1.4412 step 689( 5 secs): Minibatch rotation:angle: -0.5182 step 691( 5 secs): Minibatch rotation:angle: 0.6104 step 700( 5 secs): Minibatch rotation:angle: 0.2777 step 713( 5 secs): Minibatch rotation:angle: 0.4088 step 731( 5 secs): Minibatch rotation:angle: 1.5995 step 761( 5 secs): Minibatch rotation:angle: -0.1049 step 764( 5 secs): Minibatch rotation:angle: -0.1286 step 785( 5 secs): Minibatch rotation:angle: 0.2296 step 795( 5 secs): Minibatch rotation:angle: -1.4570 step 813( 7 secs): Minibatch rotation:angle: -1.5104 step 837( 7 secs): Minibatch rotation:angle: 0.9490 step 847( 7 secs): Minibatch rotation:angle: 1.2389 step 857( 7 secs): Minibatch rotation:angle: 1.9708 step 861( 7 secs): Minibatch rotation:angle: -1.3822 step 877( 7 secs): Minibatch rotation:angle: -1.6069 step 880( 7 secs): Minibatch rotation:angle: 1.4276 step 882( 7 secs): Minibatch rotation:angle: 0.0323 step 895( 7 secs): Minibatch rotation:angle: -0.7840 step 909( 7 secs): Minibatch rotation:angle: 0.2200 step 924( 7 secs): Minibatch rotation:angle: 0.9365 step 939( 7 secs): Minibatch rotation:angle: 0.5616 step 942( 7 secs): Minibatch rotation:angle: -1.3131 step 945( 7 secs): Minibatch rotation:angle: 0.5496 step 948( 7 secs): Minibatch rotation:angle: 1.8078 step 952( 7 secs): Minibatch rotation:angle: 1.6791 step 954( 7 secs): Minibatch rotation:angle: 1.3377 step 969( 7 secs): Minibatch rotation:angle: 0.3555 step 981( 7 secs): Minibatch rotation:angle: -0.8890 step 992( 7 secs): Minibatch rotation:angle: -0.7087 step 1003( 7 secs): Minibatch rotation:angle: -1.6007 step 1020( 7 secs): Minibatch rotation:angle: -1.0628 step 1032( 7 secs): Minibatch rotation:angle: -1.4750 step 1048( 7 secs): Minibatch rotation:angle: 1.5043 step 1057( 7 secs): Minibatch rotation:angle: 0.0029 step 1061( 7 secs): Minibatch rotation:angle: 0.2589 step 1067( 7 secs): Minibatch rotation:angle: -0.6912 step 1075( 7 secs): Minibatch rotation:angle: -0.4388 step 1078( 7 secs): Minibatch rotation:angle: -0.7461 step 1110( 7 secs): Minibatch rotation:angle: -1.0476 step 1130( 7 secs): Minibatch rotation:angle: 0.8604 step 1139( 7 secs): Minibatch rotation:angle: 0.1264 step 1141( 7 secs): Minibatch rotation:angle: -0.5500 step 1146( 7 secs): Minibatch rotation:angle: 0.8126 step 1151( 7 secs): Minibatch rotation:angle: -1.1652 step 1153( 7 secs): Minibatch rotation:angle: 0.5463 step 1180( 7 secs): Minibatch rotation:angle: 0.9390 step 1182( 7 secs): Minibatch rotation:angle: -1.8405 step 1189( 7 secs): Minibatch rotation:angle: -1.3134 step 1197( 7 secs): Minibatch rotation:angle: -0.1215 step 1204( 7 secs): Minibatch rotation:angle: -1.2543 step 1208( 7 secs): Minibatch rotation:angle: 0.7490 step 1229( 7 secs): Minibatch rotation:angle: 0.1293 step 1237( 7 secs): Minibatch rotation:angle: 0.3663 step 1266( 7 secs): Minibatch rotation:angle: -1.9999 step 1278( 7 secs): Minibatch rotation:angle: 1.8027 step 1295( 7 secs): Minibatch rotation:angle: 1.3164 step 1297( 7 secs): Minibatch rotation:angle: 0.9677 step 1299( 7 secs): Minibatch rotation:angle: 1.9622 step 1325( 7 secs): Minibatch rotation:angle: -1.9778 step 1328( 7 secs): Minibatch rotation:angle: -1.1786 step 1334( 7 secs): Minibatch rotation:angle: -0.7078 step 1336( 7 secs): Minibatch rotation:angle: -1.8895 step 1359( 7 secs): Minibatch rotation:angle: -1.6903 step 1382( 7 secs): Minibatch rotation:angle: 0.0515 step 1385( 7 secs): Minibatch rotation:angle: 0.8232 step 1387( 7 secs): Minibatch rotation:angle: 1.9280 step 1404( 7 secs): Minibatch rotation:angle: -1.7674 step 1415( 7 secs): Minibatch rotation:angle: -0.9821 step 1418( 7 secs): Minibatch rotation:angle: 0.2995 step 1434( 7 secs): Minibatch rotation:angle: 0.3138 step 1447( 7 secs): Minibatch rotation:angle: -0.1445 step 1468( 7 secs): Minibatch rotation:angle: 0.2396 step 1472( 7 secs): Minibatch rotation:angle: -1.0896 step 1480( 7 secs): Minibatch rotation:angle: 0.5134 step 1484( 7 secs): Minibatch rotation:angle: 0.6300 step 1512( 7 secs): Minibatch rotation:angle: 1.2379 step 1520( 7 secs): Minibatch rotation:angle: 0.5878 step 1532( 7 secs): Minibatch rotation:angle: 0.7924 step 1544( 7 secs): Minibatch rotation:angle: -1.5924 step 1546( 7 secs): Minibatch rotation:angle: 0.4798 step 1551( 7 secs): Minibatch rotation:angle: 0.6005 step 1569( 7 secs): Minibatch rotation:angle: -0.9395 step 1594( 7 secs): Minibatch rotation:angle: 1.6593 step 1600( 7 secs): Minibatch rotation:angle: 0.5128 step 1647( 7 secs): Minibatch rotation:angle: 1.2361 step 1663( 7 secs): Minibatch rotation:angle: -1.2202 step 1693( 7 secs): Minibatch rotation:angle: -1.6133 step 1695( 7 secs): Minibatch rotation:angle: -0.9869 step 1701( 7 secs): Minibatch rotation:angle: -0.2891 step 1709( 7 secs): Minibatch rotation:angle: 1.0592 step 1715( 7 secs): Minibatch rotation:angle: -1.7232 step 1716( 7 secs): Minibatch rotation:angle: 0.4602 step 1721( 7 secs): Minibatch rotation:angle: 1.8439 step 1723( 7 secs): Minibatch rotation:angle: 1.2296 step 1743( 7 secs): Minibatch rotation:angle: 1.4123 step 1746( 7 secs): Minibatch rotation:angle: 0.5527 step 1747( 7 secs): Minibatch rotation:angle: -1.4935 step 1768( 7 secs): Minibatch rotation:angle: -0.2607 step 1772( 7 secs): Minibatch rotation:angle: 0.7055 step 1781( 7 secs): Minibatch rotation:angle: 0.3354 step 1817( 7 secs): Minibatch rotation:angle: -1.9580 step 1821( 7 secs): Minibatch rotation:angle: -1.5412 step 1825( 7 secs): Minibatch rotation:angle: -0.4509 step 1827( 7 secs): Minibatch rotation:angle: 1.9688 step 1851( 7 secs): Minibatch rotation:angle: -1.7624 step 1857( 7 secs): Minibatch rotation:angle: 0.5838 step 1862( 7 secs): Minibatch rotation:angle: -0.3668 step 1874( 7 secs): Minibatch rotation:angle: -1.4491 step 1875( 7 secs): Minibatch rotation:angle: -0.3541 step 1888( 7 secs): Minibatch rotation:angle: 0.7662 step 1894( 7 secs): Minibatch rotation:angle: 0.4572 step 1905( 7 secs): Minibatch rotation:angle: -1.1159 step 1906( 7 secs): Minibatch rotation:angle: -1.9946 step 1915( 7 secs): Minibatch rotation:angle: 0.1886 step 1923( 7 secs): Minibatch rotation:angle: 0.5597 step 1938( 7 secs): Minibatch rotation:angle: -1.5014 step 1940( 7 secs): Minibatch rotation:angle: -0.7812 step 1941( 7 secs): Minibatch rotation:angle: -0.9926 step 1977( 7 secs): Minibatch rotation:angle: 1.9638 step 1978( 7 secs): Minibatch rotation:angle: -1.4872 step 1991( 7 secs): Minibatch rotation:angle: 1.3604 step 2008( 18 secs): Minibatch rotation:angle: -0.4081 step 2014( 18 secs): Minibatch rotation:angle: -1.6980 step 2027( 18 secs): Minibatch rotation:angle: -0.2367 step 2030( 18 secs): Minibatch rotation:angle: -0.8204 step 2041( 18 secs): Minibatch rotation:angle: -1.3729 step 2046( 18 secs): Minibatch rotation:angle: -0.2749 step 2081( 18 secs): Minibatch rotation:angle: -1.9513 step 2084( 18 secs): Minibatch rotation:angle: -1.8022 step 2098( 18 secs): Minibatch rotation:angle: 1.0991 step 2100( 18 secs): Minibatch rotation:angle: -0.5681 step 2101( 18 secs): Minibatch rotation:angle: -1.7471 step 2111( 18 secs): Minibatch rotation:angle: -1.4369 step 2132( 18 secs): Minibatch rotation:angle: -0.5205 step 2148( 18 secs): Minibatch rotation:angle: 1.5780 step 2153( 18 secs): Minibatch rotation:angle: 0.6324 step 2158( 18 secs): Minibatch rotation:angle: 1.9805 step 2159( 18 secs): Minibatch rotation:angle: 0.9734 step 2164( 18 secs): Minibatch rotation:angle: 1.1774 step 2172( 18 secs): Minibatch rotation:angle: 1.2144 step 2185( 18 secs): Minibatch rotation:angle: -0.7179 step 2203( 18 secs): Minibatch rotation:angle: -1.5074 step 2205( 18 secs): Minibatch rotation:angle: 0.3687 step 2233( 18 secs): Minibatch rotation:angle: -0.8576 step 2243( 18 secs): Minibatch rotation:angle: -0.8914 step 2249( 18 secs): Minibatch rotation:angle: 1.4322 step 2250( 18 secs): Minibatch rotation:angle: 1.5765 step 2258( 18 secs): Minibatch rotation:angle: -1.7482 step 2259( 18 secs): Minibatch rotation:angle: 1.9248 step 2265( 18 secs): Minibatch rotation:angle: -0.2367 step 2322( 18 secs): Minibatch rotation:angle: -0.1462 step 2334( 18 secs): Minibatch rotation:angle: -1.2451 step 2343( 18 secs): Minibatch rotation:angle: -1.7421 step 2344( 18 secs): Minibatch rotation:angle: 0.7753 step 2345( 18 secs): Minibatch rotation:angle: -1.0354 step 2347( 18 secs): Minibatch rotation:angle: 1.0084 step 2350( 18 secs): Minibatch rotation:angle: -0.0563 step 2355( 18 secs): Minibatch rotation:angle: 1.5616 step 2371( 18 secs): Minibatch rotation:angle: 0.1832 step 2420( 18 secs): Minibatch rotation:angle: 0.7866 step 2423( 18 secs): Minibatch rotation:angle: -1.6549 step 2424( 18 secs): Minibatch rotation:angle: 0.6501 step 2436( 18 secs): Minibatch rotation:angle: -1.9990 step 2441( 18 secs): Minibatch rotation:angle: 1.3462 step 2470( 18 secs): Minibatch rotation:angle: 1.1106 step 2472( 18 secs): Minibatch rotation:angle: -1.1674 step 2474( 18 secs): Minibatch rotation:angle: 1.3391 step 2486( 18 secs): Minibatch rotation:angle: -1.1047 step 2505( 18 secs): Minibatch rotation:angle: -1.2807 step 2520( 18 secs): Minibatch rotation:angle: 1.4115 step 2524( 18 secs): Minibatch rotation:angle: -1.5249 step 2537( 18 secs): Minibatch rotation:angle: -0.5955 step 2545( 18 secs): Minibatch rotation:angle: 0.4093 step 2572( 18 secs): Minibatch rotation:angle: -1.3984 step 2573( 18 secs): Minibatch rotation:angle: 1.7442 step 2576( 18 secs): Minibatch rotation:angle: -1.2091 step 2592( 18 secs): Minibatch rotation:angle: -0.3338 step 2594( 18 secs): Minibatch rotation:angle: 0.5814 step 2609( 18 secs): Minibatch rotation:angle: 0.4164 step 2611( 18 secs): Minibatch rotation:angle: -1.0385 step 2639( 18 secs): Minibatch rotation:angle: -0.3355 step 2640( 18 secs): Minibatch rotation:angle: 0.6962 step 2642( 18 secs): Minibatch rotation:angle: -0.5901 step 2653( 18 secs): Minibatch rotation:angle: 1.9364 step 2667( 18 secs): Minibatch rotation:angle: -0.9125 step 2669( 18 secs): Minibatch rotation:angle: -1.4134 step 2670( 18 secs): Minibatch rotation:angle: -0.4466 step 2683( 18 secs): Minibatch rotation:angle: 1.3829 step 2685( 18 secs): Minibatch rotation:angle: 0.2787 step 2690( 18 secs): Minibatch rotation:angle: -1.5205 step 2740( 18 secs): Minibatch rotation:angle: 0.1628 step 2741( 18 secs): Minibatch rotation:angle: 0.3390 step 2746( 18 secs): Minibatch rotation:angle: -0.3084 step 2750( 18 secs): Minibatch rotation:angle: 1.9089 step 2755( 18 secs): Minibatch rotation:angle: -0.3149 step 2758( 18 secs): Minibatch rotation:angle: 1.7752 step 2764( 18 secs): Minibatch rotation:angle: 1.7116 step 2765( 18 secs): Minibatch rotation:angle: -1.5599 step 2780( 18 secs): Minibatch rotation:angle: -1.0074 step 2782( 18 secs): Minibatch rotation:angle: -1.2068 step 2791( 18 secs): Minibatch rotation:angle: 1.9188 step 2800( 18 secs): Minibatch rotation:angle: 0.4213 step 2803( 18 secs): Minibatch rotation:angle: -0.2262 Vld accuracy:0.8332 [ 0.7045 0.91 0.7876 0.7314 1. 0.9408 0.8781 0.9914 0.3379 0.9978] [[341 0 0 1 113 2 0 3 0 24] [ 0 435 0 0 20 2 1 15 0 5] [ 0 0 356 0 32 0 33 25 0 6] [ 0 0 0 324 119 0 0 0 0 0] [ 0 0 0 0 459 0 0 0 0 0] [ 1 0 0 0 17 397 1 0 0 6] [ 0 0 0 0 44 0 389 5 0 5] [ 0 0 0 0 2 0 1 347 0 0] [ 1 1 1 0 89 0 50 71 123 28] [ 0 0 0 0 0 0 0 1 0 451]]
Vld logLoss:4.8369 [ 0.9739 0.2233 0.6196 0.6802 0. 0.1763 0.3445 0.0209 1.7902 0.008 ] max Pby for cls: c0; desc: normal driving; proba: 1.0000; nObs: 320 img_62023.jpg: plot_occlusion:
display_weight:
Proba: [ 1. 0. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c0; desc: normal driving; proba: 0.6795; nObs: 1 img_80584.jpg: plot_occlusion:
display_weight:
Proba: [ 0.6795 0. 0. 0. 0. 0. 0. 0. 0. 0.3205] next best class: talking to passenger max Pby for cls: c1; desc: texting - right; proba: 1.0000; nObs: 426 img_89322.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c1; desc: texting - right; proba: 0.8060; nObs: 1 img_83254.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.806 0. 0. 0. 0.194 0. 0. 0. 0. ] next best class: operating the radio max Pby for cls: c2; desc: talking on the phone - right; proba: 1.0000; nObs: 342 img_87211.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c2; desc: talking on the phone - right; proba: 0.6004; nObs: 1 img_53195.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.6004 0. 0.3996 0. 0. 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c3; desc: texting - left; proba: 1.0000; nObs: 288 img_77361.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.] min Pby for cls: c3; desc: texting - left; proba: 0.5175; nObs: 1 img_94042.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.5175 0.4825 0. 0. 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c4; desc: talking on the phone - left; proba: 1.0000; nObs: 807 img_78749.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.] min Pby for cls: c4; desc: talking on the phone - left; proba: 0.5045; nObs: 1 img_12442.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.4955 0.5045 0. 0. 0. 0. 0. ] next best class: texting - left max Pby for cls: c5; desc: operating the radio; proba: 1.0000; nObs: 398 img_70461.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.] min Pby for cls: c5; desc: operating the radio; proba: 0.8562; nObs: 1 img_54029.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.1438 0. 0. 0. 0.8562 0. 0. 0. 0. ] next best class: texting - right max Pby for cls: c6; desc: drinking; proba: 1.0000; nObs: 440 img_68722.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.] min Pby for cls: c6; desc: drinking; proba: 0.7608; nObs: 1 img_32199.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0.2392 0. 0.7608 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c7; desc: reaching behind; proba: 1.0000; nObs: 433 img_29251.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.] min Pby for cls: c7; desc: reaching behind; proba: 0.4930; nObs: 1 img_88262.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.0498 0. 0. 0. 0.4572 0.493 0. 0. ] next best class: drinking max Pby for cls: c8; desc: hair and makeup; proba: 1.0000; nObs: 107 img_8759.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.] min Pby for cls: c8; desc: hair and makeup; proba: 0.8165; nObs: 1 img_9578.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0.1835 0. 0. 0. 0.8165 0. ] next best class: talking on the phone - left max Pby for cls: c9; desc: talking to passenger; proba: 1.0000; nObs: 497 img_24934.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.] min Pby for cls: c9; desc: talking to passenger; proba: 0.8207; nObs: 1 img_60533.jpg: plot_occlusion:
display_weight:
Proba:
[ 0. 0. 0. 0. 0.1793 0. 0. 0. 0.
0.8207]
next best class: talking on the phone - left
predicting 79726 new obs...
@ 68 secs: obsIx: 0
@ 69 secs: obsIx: 8000
@ 74 secs: obsIx: 40000
max Pby for cls: c0; desc: normal driving; proba: 1.0000; nObs: 3083
img_100.jpg:
plot_occlusion:
display_weight:
Proba: [ 1. 0. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c0; desc: normal driving; proba: 0.5038; nObs: 1 img_91731.jpg: plot_occlusion:
display_weight:
Proba: [ 0.5038 0. 0. 0. 0.4962 0. 0. 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c1; desc: texting - right; proba: 1.0000; nObs: 3904 img_100003.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 1. 0. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c1; desc: texting - right; proba: 0.4488; nObs: 1 img_58091.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0.4488 0. 0. 0.2213 0. 0. 0. 0. 0.3298] next best class: talking to passenger max Pby for cls: c2; desc: talking on the phone - right; proba: 1.0000; nObs: 1733 img_10001.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 1. 0. 0. 0. 0. 0. 0. 0.] min Pby for cls: c2; desc: talking on the phone - right; proba: 0.5096; nObs: 1 img_57784.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0.5096 0. 0. 0. 0. 0. 0. 0.4904] next best class: talking to passenger max Pby for cls: c3; desc: texting - left; proba: 1.0000; nObs: 3918 img_10.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 1. 0. 0. 0. 0. 0. 0.] min Pby for cls: c3; desc: texting - left; proba: 0.5045; nObs: 1 img_78739.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.5045 0.0001 0. 0. 0. 0. 0.4955] next best class: talking to passenger max Pby for cls: c4; desc: talking on the phone - left; proba: 1.0000; nObs: 24812 img_1000.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 1. 0. 0. 0. 0. 0.] min Pby for cls: c4; desc: talking on the phone - left; proba: 0.4816; nObs: 1 img_68076.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0.4636 0.4816 0. 0. 0. 0. 0.0547] next best class: texting - left max Pby for cls: c5; desc: operating the radio; proba: 1.0000; nObs: 5699 img_100001.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 1. 0. 0. 0. 0.] min Pby for cls: c5; desc: operating the radio; proba: 0.3664; nObs: 1 img_78625.jpg: plot_occlusion:
display_weight:
Proba: [ 0.2851 0. 0. 0. 0.3485 0.3664 0. 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c6; desc: drinking; proba: 1.0000; nObs: 5312 img_100004.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 1. 0. 0. 0.] min Pby for cls: c6; desc: drinking; proba: 0.5082; nObs: 1 img_48373.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0.4918 0. 0.5082 0. 0. 0. ] next best class: talking on the phone - left max Pby for cls: c7; desc: reaching behind; proba: 1.0000; nObs: 6124 img_100005.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 1. 0. 0.] min Pby for cls: c7; desc: reaching behind; proba: 0.5014; nObs: 1 img_22445.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0.4986 0. 0. 0.5014 0. 0. ] next best class: talking on the phone - left max Pby for cls: c8; desc: hair and makeup; proba: 1.0000; nObs: 4040 img_100035.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 1. 0.] min Pby for cls: c8; desc: hair and makeup; proba: 0.5028; nObs: 1 img_70747.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 0.5028 0.4972] next best class: talking to passenger max Pby for cls: c9; desc: talking to passenger; proba: 1.0000; nObs: 11551 img_100024.jpg: plot_occlusion:
display_weight:
Proba: [ 0. 0. 0. 0. 0. 0. 0. 0. 0. 1.] min Pby for cls: c9; desc: talking to passenger; proba: 0.4451; nObs: 1 img_89849.jpg: plot_occlusion:
display_weight:
Proba:
[ 0. 0. 0.2253 0. 0. 0. 0. 0.3296 0.
0.4451]
next best class: reaching behind
New prediction knts:
{'kntCls': (array([0, 1, 2, 3, 4, 5, 6, 7, 8, 9]), array([ 3619, 4408, 2109, 4576, 27385, 6501, 6198, 6814, 4881, 13235]))}
duration: 122 seconds
In [25]:
print finMdlDf
id lrnRateTfw nObsBtc nObsFit nStepsTfw rotateMaxAgl \
0 LgtRgr.SGD.tfw 3.0 8.0 22424 2804.0 2
rotatePby accVld accVldCls \
0 0.1 0.833218 {u'accCls': [0.704545454545, 0.910041841004, 0...
logLossVld logLossVldCls \
0 4.836929 {u'logLossCls': [0.973879753517, 0.22334474905...
predNew \
0 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
model duration
0 <tensorflow.python.client.session.Session obje... 122
In [26]:
glbMdlDf = glbMdlDf.append(finMdlDf)
glbMdlDf = glbMdlDf.set_index(['id'] + srchParamsDct.keys(), drop = False)
glbMdlDf = glbMdlDf.sort_values(
['nObsFit', 'accVld', 'logLossVld', 'duration'],
ascending = [False , True , False, False])
print(glbMdlDf[list(set(glbMdlDf.columns) -
set(['id'] + srchParamsDct.keys()))])
accVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 {u'accCls': [0.704545454545, 0.910041841004, 0...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'accCls': [0.989669421488, 0.991631799163, 0...
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 {u'accCls': [0.297520661157, 0.081589958159, 0...
100.0 18077.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
1000.0 18077.0 0.0 16.0 2.0 10 {u'accCls': [0.495867768595, 0.0397489539749, ...
1 {u'accCls': [0.495867768595, 0.0397489539749, ...
0 {u'accCls': [0.495867768595, 0.0397489539749, ...
2 {u'accCls': [0.495867768595, 0.0397489539749, ...
0.1 16.0 3.0 10 {u'accCls': [0.528925619835, 0.0523012552301, ...
4.0 3.0 1 {u'accCls': [0.545454545455, 0.0690376569038, ...
8.0 2.0 0 {u'accCls': [0.132231404959, 0.010460251046, 0...
0.0 8.0 2.0 10 {u'accCls': [0.132231404959, 0.010460251046, 0...
1 {u'accCls': [0.132231404959, 0.010460251046, 0...
0 {u'accCls': [0.132231404959, 0.010460251046, 0...
2 {u'accCls': [0.132231404959, 0.010460251046, 0...
0.5 4.0 2.0 10 {u'accCls': [0.328512396694, 0.092050209205, 0...
8.0 3.0 10 {u'accCls': [0.0433884297521, 0.202928870293, ...
16.0 3.0 10 {u'accCls': [0.386363636364, 0.119246861925, 0...
0.0 8.0 10.0 10 {u'accCls': [0.51652892562, 0.0292887029289, 0...
0.1 16.0 1.0 0 {u'accCls': [0.340909090909, 0.169456066946, 0...
0.0 16.0 1.0 10 {u'accCls': [0.340909090909, 0.169456066946, 0...
1 {u'accCls': [0.340909090909, 0.169456066946, 0...
2 {u'accCls': [0.340909090909, 0.169456066946, 0...
0 {u'accCls': [0.340909090909, 0.169456066946, 0...
0.1 4.0 2.0 2 {u'accCls': [0.564049586777, 0.673640167364, 0...
8.0 2.0 3 {u'accCls': [0.202479338843, 0.117154811715, 0...
16.0 2.0 10 {u'accCls': [0.252066115702, 0.10460251046, 0....
0.0 8.0 4.0 10 {u'accCls': [0.229338842975, 0.00418410041841,...
1 {u'accCls': [0.229338842975, 0.00418410041841,...
... ...
0.1 8.0 3.0 0 {u'accCls': [0.460743801653, 0.640167364017, 0...
0.0 8.0 3.0 1 {u'accCls': [0.460743801653, 0.640167364017, 0...
0 {u'accCls': [0.460743801653, 0.640167364017, 0...
2 {u'accCls': [0.460743801653, 0.640167364017, 0...
0.1 8.0 2.0 1 {u'accCls': [0.338842975207, 0.347280334728, 0...
1.0 2 {u'accCls': [0.448347107438, 0.44769874477, 0....
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'accCls': [0.504132231405, 0.435146443515, 0...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'accCls': [0.0, 0.150627615063, 0.0, 0.00225...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 {u'accCls': [0.169421487603, 0.144351464435, 0...
1.0 0 {u'accCls': [0.289256198347, 0.182008368201, 0...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'accCls': [0.456611570248, 0.380753138075, 0...
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 {u'accCls': [0.607438016529, 0.219665271967, 0...
8000.0 0.1 8.0 3.0 2 {u'accCls': [0.553719008264, 0.610878661088, 0...
7000.0 0.1 8.0 3.0 2 {u'accCls': [0.431818181818, 0.558577405858, 0...
6000.0 0.1 8.0 3.0 2 {u'accCls': [0.181818181818, 0.671548117155, 0...
5000.0 0.1 8.0 3.0 2 {u'accCls': [0.543388429752, 0.535564853556, 0...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
2000.0 0.0 2000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'accCls': [0.150826446281, 0.142259414226, 0...
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 {u'accCls': [0.136363636364, 0.0230125523013, ...
1000.0 1000.0 0.1 8.0 3.0 2 {u'accCls': [0.204545454545, 0.0334728033473, ...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 {u'accCls': [0.423553719008, 0.182008368201, 0...
10.0 0 {u'accCls': [0.646694214876, 0.0564853556485, ...
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'accCls': [0.0185950413223, 0.133891213389, ...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
bestFit \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 False
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 False
100.0 18077.0 0.1 16.0 1.0 10 False
1000.0 18077.0 0.0 16.0 2.0 10 False
1 False
0 False
2 False
0.1 16.0 3.0 10 False
4.0 3.0 1 False
8.0 2.0 0 False
0.0 8.0 2.0 10 False
1 False
0 False
2 False
0.5 4.0 2.0 10 False
8.0 3.0 10 False
16.0 3.0 10 False
0.0 8.0 10.0 10 False
0.1 16.0 1.0 0 False
0.0 16.0 1.0 10 False
1 False
2 False
0 False
0.1 4.0 2.0 2 False
8.0 2.0 3 False
16.0 2.0 10 False
0.0 8.0 4.0 10 False
1 False
... ...
0.1 8.0 3.0 0 False
0.0 8.0 3.0 1 False
0 False
2 False
0.1 8.0 2.0 1 False
1.0 2 False
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 True
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 False
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 False
1.0 0 False
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 False
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 False
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 False
8000.0 0.1 8.0 3.0 2 False
7000.0 0.1 8.0 3.0 2 False
6000.0 0.1 8.0 3.0 2 False
5000.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 False
2000.0 0.0 2000.0 -1.0 0 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 False
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 False
1000.0 1000.0 0.1 8.0 3.0 2 False
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 False
10.0 0 False
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 False
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 False
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 False
logLossVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 4.836929
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.193149
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 0.018463
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 26.843734
100.0 18077.0 0.1 16.0 1.0 10 24.085776
1000.0 18077.0 0.0 16.0 2.0 10 25.605293
1 25.605293
0 25.605293
2 25.605293
0.1 16.0 3.0 10 25.966046
4.0 3.0 1 26.212732
8.0 2.0 0 24.909087
0.0 8.0 2.0 10 24.909087
1 24.909087
0 24.909087
2 24.909087
0.5 4.0 2.0 10 24.777048
8.0 3.0 10 24.417234
16.0 3.0 10 24.161626
0.0 8.0 10.0 10 24.927648
0.1 16.0 1.0 0 20.958448
0.0 16.0 1.0 10 20.958448
1 20.958448
2 20.958448
0 20.958448
0.1 4.0 2.0 2 24.175734
8.0 2.0 3 23.044421
16.0 2.0 10 22.249012
0.0 8.0 4.0 10 24.110602
1 24.110602
... ...
0.1 8.0 3.0 0 20.948740
0.0 8.0 3.0 1 20.948740
0 20.948740
2 20.948740
0.1 8.0 2.0 1 20.091985
1.0 2 17.898932
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 19.209913
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 20.546733
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 3.085191
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 24.085776
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 8.912652
1.0 0 9.502959
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 2.885114
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 20.546733
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 18.995890
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 20.546733
8000.0 0.1 8.0 3.0 2 20.345963
7000.0 0.1 8.0 3.0 2 20.307083
6000.0 0.1 8.0 3.0 2 21.354761
5000.0 0.1 8.0 3.0 2 21.899301
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 2.610439
2000.0 0.0 2000.0 -1.0 0 2.498749
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 12.600104
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 18.141110
1000.0 1000.0 0.1 8.0 3.0 2 25.728109
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 14.714048
10.0 0 24.274683
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 2.539650
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 25.837055
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 2.497940
predNew \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'kntCls': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.0 16.0 2.0 10 NaN
1 NaN
0 NaN
2 NaN
0.1 16.0 3.0 10 NaN
4.0 3.0 1 NaN
8.0 2.0 0 NaN
0.0 8.0 2.0 10 NaN
1 NaN
0 NaN
2 NaN
0.5 4.0 2.0 10 NaN
8.0 3.0 10 NaN
16.0 3.0 10 NaN
0.0 8.0 10.0 10 NaN
0.1 16.0 1.0 0 NaN
0.0 16.0 1.0 10 NaN
1 NaN
2 NaN
0 NaN
0.1 4.0 2.0 2 NaN
8.0 2.0 3 NaN
16.0 2.0 10 NaN
0.0 8.0 4.0 10 NaN
1 NaN
... ...
0.1 8.0 3.0 0 NaN
0.0 8.0 3.0 1 NaN
0 NaN
2 NaN
0.1 8.0 2.0 1 NaN
1.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 NaN
1.0 0 NaN
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
5000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
2000.0 0.0 2000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 NaN
1000.0 1000.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 NaN
10.0 0 NaN
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 {u'clsKnt': ([0, 1, 2, 3, 4, 5, 6, 7, 8, 9], [...
logLossVldCls \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 {u'logLossCls': [0.973879753517, 0.22334474905...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 {u'logLossCls': [0.0123335882818, 0.0241534716...
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 {u'logLossCls': [2.62356084168, 3.42588099594,...
100.0 18077.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
1000.0 18077.0 0.0 16.0 2.0 10 {u'logLossCls': [1.86638576386, 3.42782085718,...
1 {u'logLossCls': [1.86638576386, 3.42782085718,...
0 {u'logLossCls': [1.86638576386, 3.42782085718,...
2 {u'logLossCls': [1.86638576386, 3.42782085718,...
0.1 16.0 3.0 10 {u'logLossCls': [1.71310413171, 3.4687167196, ...
4.0 3.0 1 {u'logLossCls': [1.71708309944, 3.4968377777, ...
8.0 2.0 0 {u'logLossCls': [3.05751075809, 3.73295341141,...
0.0 8.0 2.0 10 {u'logLossCls': [3.05751075809, 3.73295341141,...
1 {u'logLossCls': [3.05751075809, 3.73295341141,...
0 {u'logLossCls': [3.05751075809, 3.73295341141,...
2 {u'logLossCls': [3.05751075809, 3.73295341141,...
0.5 4.0 2.0 10 {u'logLossCls': [2.38429593137, 3.32281736573,...
8.0 3.0 10 {u'logLossCls': [3.61623259217, 2.83198596779,...
16.0 3.0 10 {u'logLossCls': [2.15663510596, 3.1841831293, ...
0.0 8.0 10.0 10 {u'logLossCls': [1.82418636316, 3.66306265443,...
0.1 16.0 1.0 0 {u'logLossCls': [2.09243845599, 2.30268346691,...
0.0 16.0 1.0 10 {u'logLossCls': [2.09243845599, 2.30268346691,...
1 {u'logLossCls': [2.09243845599, 2.30268346691,...
2 {u'logLossCls': [2.09243845599, 2.30268346691,...
0 {u'logLossCls': [2.09243845599, 2.30268346691,...
0.1 4.0 2.0 2 {u'logLossCls': [1.59263670019, 1.19242519276,...
8.0 2.0 3 {u'logLossCls': [2.7750707387, 3.1253805868, 3...
16.0 2.0 10 {u'logLossCls': [2.42199459643, 3.05959930614,...
0.0 8.0 4.0 10 {u'logLossCls': [2.81500682324, 3.7814798951, ...
1 {u'logLossCls': [2.81500682324, 3.7814798951, ...
... ...
0.1 8.0 3.0 0 {u'logLossCls': [1.99287215213, 1.25414623142,...
0.0 8.0 3.0 1 {u'logLossCls': [1.99287215213, 1.25414623142,...
0 {u'logLossCls': [1.99287215213, 1.25414623142,...
2 {u'logLossCls': [1.99287215213, 1.25414623142,...
0.1 8.0 2.0 1 {u'logLossCls': [2.4098162987, 2.1249522721, 2...
1.0 2 {u'logLossCls': [1.7477126908, 1.45845704706, ...
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 {u'logLossCls': [1.72596805748, 1.7853875361, ...
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 {u'logLossCls': [3.84558724986, 2.69565642157,...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 {u'logLossCls': [1.11536194976, 1.25718941188,...
1.0 0 {u'logLossCls': [1.14059053163, 1.46723165909,...
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 {u'logLossCls': [1.84540699409, 1.93896956479,...
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.48863261945, 2.76463156066,...
8000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.5257602538, 1.31972597838, ...
7000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.06754908637, 1.51088086827,...
6000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.97387028973, 1.20522205183,...
5000.0 0.1 8.0 3.0 2 {u'logLossCls': [1.57533202071, 1.59023941247,...
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
2000.0 0.0 2000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 {u'logLossCls': [1.47411393905, 2.08387227918,...
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 {u'logLossCls': [2.1826821181, 3.34540505739, ...
1000.0 1000.0 0.1 8.0 3.0 2 {u'logLossCls': [2.84808429073, 3.58756550329,...
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 {u'logLossCls': [0.842354858019, 1.17531354279...
10.0 0 {u'logLossCls': [1.28781316316, 3.51519487621,...
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 {u'logLossCls': [3.67759601248, 3.22267129616,...
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
accVld \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 0.833218
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 0.987348
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 1.000000
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 0.208420
100.0 18077.0 0.1 16.0 1.0 10 0.210950
1000.0 18077.0 0.0 16.0 2.0 10 0.214861
1 0.214861
0 0.214861
2 0.214861
0.1 16.0 3.0 10 0.219692
4.0 3.0 1 0.226823
8.0 2.0 0 0.241776
0.0 8.0 2.0 10 0.241776
1 0.241776
0 0.241776
2 0.241776
0.5 4.0 2.0 10 0.252588
8.0 3.0 10 0.262020
16.0 3.0 10 0.268231
0.0 8.0 10.0 10 0.269841
0.1 16.0 1.0 0 0.270761
0.0 16.0 1.0 10 0.270761
1 0.270761
2 0.270761
0 0.270761
0.1 4.0 2.0 2 0.272602
8.0 2.0 3 0.272832
16.0 2.0 10 0.273522
0.0 8.0 4.0 10 0.276282
1 0.276282
... ...
0.1 8.0 3.0 0 0.357028
0.0 8.0 3.0 1 0.357028
0 0.357028
2 0.357028
0.1 8.0 2.0 1 0.357488
1.0 2 0.361859
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 0.375201
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 0.381643
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 0.342535
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 0.210950
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 0.251208
1.0 0 0.319991
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 0.358638
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 0.381643
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 0.384173
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 0.381643
8000.0 0.1 8.0 3.0 2 0.378882
7000.0 0.1 8.0 3.0 2 0.375431
6000.0 0.1 8.0 3.0 2 0.339545
5000.0 0.1 8.0 3.0 2 0.320911
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 0.363699
2000.0 0.0 2000.0 -1.0 0 0.341385
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 0.173453
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 0.212790
1000.0 1000.0 0.1 8.0 3.0 2 0.224983
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 0.255809
10.0 0 0.273522
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 0.333333
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 0.224753
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 0.307108
duration \
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 122
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 743
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 874
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 12
100.0 18077.0 0.1 16.0 1.0 10 2
1000.0 18077.0 0.0 16.0 2.0 10 11
1 11
0 11
2 10
0.1 16.0 3.0 10 13
4.0 3.0 1 9
8.0 2.0 0 10
0.0 8.0 2.0 10 9
1 9
0 8
2 8
0.5 4.0 2.0 10 11
8.0 3.0 10 16
16.0 3.0 10 23
0.0 8.0 10.0 10 9
0.1 16.0 1.0 0 70
0.0 16.0 1.0 10 12
1 11
2 11
0 10
0.1 4.0 2.0 2 8
8.0 2.0 3 10
16.0 2.0 10 13
0.0 8.0 4.0 10 9
1 9
... ...
0.1 8.0 3.0 0 10
0.0 8.0 3.0 1 9
0 8
2 8
0.1 8.0 2.0 1 10
1.0 2 9
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 519
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 11
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 514
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 2
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 295
1.0 0 301
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 317
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 9
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 288
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 9
8000.0 0.1 8.0 3.0 2 9
7000.0 0.1 8.0 3.0 2 9
6000.0 0.1 8.0 3.0 2 9
5000.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 150
2000.0 0.0 2000.0 -1.0 0 55
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 5
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 2
1000.0 1000.0 0.1 8.0 3.0 2 9
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 5
10.0 0 7
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 28
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 9
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 8
model
id nStepsTfw nObsFit rotatePby nObsBtc lrnRateTfw rotateMaxAgl
LgtRgr.SGD.tfw 2804.0 22424.0 0.1 8.0 3.0 2 <tensorflow.python.client.session.Session obje...
LgtRgr.tfw 1000.0 22424.0 0.0 22424.0 10.0 0 NaN
LgtRgr.skl -1.0 22424.0 0.0 22424.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.5 4.0 3.0 10 NaN
100.0 18077.0 0.1 16.0 1.0 10 NaN
1000.0 18077.0 0.0 16.0 2.0 10 NaN
1 NaN
0 NaN
2 NaN
0.1 16.0 3.0 10 NaN
4.0 3.0 1 NaN
8.0 2.0 0 NaN
0.0 8.0 2.0 10 NaN
1 NaN
0 NaN
2 NaN
0.5 4.0 2.0 10 NaN
8.0 3.0 10 NaN
16.0 3.0 10 NaN
0.0 8.0 10.0 10 NaN
0.1 16.0 1.0 0 NaN
0.0 16.0 1.0 10 NaN
1 NaN
2 NaN
0 NaN
0.1 4.0 2.0 2 NaN
8.0 2.0 3 NaN
16.0 2.0 10 NaN
0.0 8.0 4.0 10 NaN
1 NaN
... ...
0.1 8.0 3.0 0 NaN
0.0 8.0 3.0 1 NaN
0 NaN
2 NaN
0.1 8.0 2.0 1 NaN
1.0 2 NaN
LgtRgr.tfw 1000.0 18077.0 0.0 18077.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 18077.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 15000.0 0.0 15000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 100.0 10000.0 0.1 16.0 1.0 10 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 0.1 0 NaN
1.0 0 NaN
LgtRgr.skl -1.0 10000.0 0.0 10000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 10000.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 1000.0 10000.0 0.0 10000.0 10.0 0 NaN
LgtRgr.SGD.tfw 1000.0 9000.0 0.1 8.0 3.0 2 NaN
8000.0 0.1 8.0 3.0 2 NaN
7000.0 0.1 8.0 3.0 2 NaN
6000.0 0.1 8.0 3.0 2 NaN
5000.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 5000.0 0.0 5000.0 -1.0 0 NaN
2000.0 0.0 2000.0 -1.0 0 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 0.1 0 NaN
LgtRgr.SGD.tfw 100.0 1000.0 0.1 16.0 1.0 10 NaN
1000.0 1000.0 0.1 8.0 3.0 2 NaN
LgtRgr.tfw 100.0 1000.0 0.0 1000.0 1.0 0 NaN
10.0 0 NaN
LgtRgr.skl -1.0 1000.0 0.0 1000.0 -1.0 0 NaN
LgtRgr.SGD.tfw 1000.0 100.0 0.1 8.0 3.0 2 NaN
LgtRgr.skl -1.0 100.0 0.0 100.0 -1.0 0 NaN
[121 rows x 8 columns]
In [27]:
myexportDf(glbMdlDf,
save_filepathname = glbPickleFile['models'],
save_drop_cols = 'model'
)
Compressed pickle file: data/img_M_SFDD_ImgSz_64.pickle; size: 41 KB
Output submission¶
In [28]:
print '\n finMdl:'
glbwriteSubmission(glbINew, finYNewPby,
'img_04_fit_lgtRgr_SGD_Tfw_SFDD_ImgSz_' + str(glbImg['size']) + \
'_sbmt_fin.csv')
finMdl:
img c0 c1 c2 c3 c4 c5 \
img
img_1.jpg img_1.jpg 0.0 0.0 0.0 0.0 7.011857e-02 0.0
img_10.jpg img_10.jpg 0.0 0.0 0.0 1.0 0.000000e+00 0.0
img_100.jpg img_100.jpg 1.0 0.0 0.0 0.0 1.445693e-27 0.0
img_1000.jpg img_1000.jpg 0.0 0.0 0.0 0.0 1.000000e+00 0.0
img_100000.jpg img_100000.jpg 0.0 0.0 0.0 0.0 1.000000e+00 0.0
c6 c7 c8 c9
img
img_1.jpg 0.000000e+00 0.929881 0.0 0.0
img_10.jpg 0.000000e+00 0.000000 0.0 0.0
img_100.jpg 0.000000e+00 0.000000 0.0 0.0
img_1000.jpg 6.692277e-32 0.000000 0.0 0.0
img_100000.jpg 0.000000e+00 0.000000 0.0 0.0
img c0 c1 c2 c3 c4 c5 c6 c7 c8 c9
img
img_99994.jpg img_99994.jpg 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0
img_99995.jpg img_99995.jpg 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0
img_99996.jpg img_99996.jpg 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0
img_99998.jpg img_99998.jpg 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0
img_99999.jpg img_99999.jpg 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0
exporting 79726 rows to img_04_fit_lgtRgr_SGD_Tfw_SFDD_ImgSz_64_sbmt_fin.csv...
In [29]:
prtStr = 'LeaderBoard metric for this sel submission: %0.5f vs. ' + \
'logLossVld (sel): %0.5f'
print prtStr % (22.62562, 20.5467)
prtStr = 'LeaderBoard metric for this fin submission: %0.5f vs. ' + \
'logLossVld (fin): %0.5f'
print prtStr % (23.71528, 4.8369)
print 'Best score yet:%s: %0.5f' % \
('img_02_fit_lgtRgr(Skl)_SFDD_(ImgSz_32_)sbmt(_fin).csv', 2.63892)
LeaderBoard metric for this sel submission: 22.62562 vs. logLossVld (sel): 20.54670 LeaderBoard metric for this fin submission: 23.71528 vs. logLossVld (fin): 4.83690 Best score yet:img_02_fit_lgtRgr(Skl)_SFDD_(ImgSz_32_)sbmt(_fin).csv: 2.63892
Stop here¶
Following code should be in img04_fit_lgtRgrSGDTf
Let's now switch to stochastic gradient descent training instead, which is much faster.
The graph will be similar, except that instead of holding all the training data into a constant node, we create a Placeholder node which will be fed actual data at every call of sesion.run().
In [6]:
import pandas as pd
models = pd.DataFrame({'nRELUs': [0]})
#models.ix[0, 'accuracy_scoreTest'] = 0
print models
nRELUs 0 0
In [7]:
batch_size = 128
graph = tf.Graph()
with graph.as_default():
# Input data. For the training data, we use a placeholder that will be fed
# at run time with a training minibatch.
tfwObsFitFtr = tf.placeholder(tf.float32,
shape=(batch_size, glbImg['size'] * glbImg['size']))
tfwObsFitRsp = tf.placeholder(tf.float32, shape=(batch_size, glbRspClassN))
tfwObsVldFtr = tf.constant(glbXVld)
tfwObsNewFtr = tf.constant(glbXNew)
# Variables.
tfwW = tf.Variable(
tf.truncated_normal([glbImg['size'] * glbImg['size'], glbRspClassN]))
tfwB = tf.Variable(tf.zeros([glbRspClassN]))
print(tfwW.initialized_value())
print(tfwB.initialized_value())
# Training computation.
logits = tf.matmul(tfwObsFitFtr, tfwW) + tfwB
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(logits, tfwObsFitRsp))
# Optimizer.
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# Predictions for the training, validation, and test data.
tfwObsTrnPred = tf.nn.softmax(logits)
tfwObsVldPred = tf.nn.softmax(
tf.matmul(tfwObsVldFtr, tfwW) + tfwB)
tfwObsNewPred = tf.nn.softmax(tf.matmul(tfwObsNewFtr, tfwW) + tfwB)
Tensor("Identity:0", shape=TensorShape([Dimension(784), Dimension(10)]), dtype=float32)
Tensor("Identity_1:0", shape=TensorShape([Dimension(10)]), dtype=float32)
Let's run it:
In [8]:
nStepsTfw = 3001
with tf.Session(graph=graph) as session:
tf.initialize_all_variables().run()
print("Initialized")
for step in range(nStepsTfw):
# Pick an offset within the training data, which has been randomized.
# Note: we could use better randomization across epochs.
offset = (step * batch_size) % (glbYFit.shape[0] - batch_size)
# Generate a minibatch.
batch_data = glbXFit[offset:(offset + batch_size), :]
batch_labels = glbYFit[offset:(offset + batch_size), :]
# Prepare a dictionary telling the session where to feed the minibatch.
# The key of the dictionary is the placeholder node of the graph to be fed,
# and the value is the numpy array to feed to it.
feed_dict = {tfwObsFitFtr : batch_data, tfwObsFitRsp : batch_labels}
_, l, predictions = session.run(
[optimizer, loss, tfwObsTrnPred], feed_dict=feed_dict)
if (step % 500 == 0):
print("Minibatch loss at step %d: %f" % (step, l))
print("Minibatch accuracy_score: %.1f%%" % accuracy_score(predictions, batch_labels))
print("Validation accuracy_score: %.1f%%" % accuracy_score(
tfwObsVldPred.eval(), glbYVld))
print("Test accuracy_score: %.1f%%" % accuracy_score(tfwObsNewPred.eval(), glbYNew))
models.ix[0, 'accuracy_scoreVld'] = accuracy_score(tfwObsVldPred.eval(), glbYVld)
models.ix[0, 'accuracy_scoreTst'] = accuracy_score( tfwObsNewPred.eval(), glbYNew)
Initialized Minibatch loss at step 0: 17.272371 Minibatch accuracy: 6.2% Validation accuracy: 13.0% Minibatch loss at step 500: 1.435902 Minibatch accuracy: 76.6% Validation accuracy: 75.2% Minibatch loss at step 1000: 1.280029 Minibatch accuracy: 78.1% Validation accuracy: 77.2% Minibatch loss at step 1500: 1.147653 Minibatch accuracy: 77.3% Validation accuracy: 77.2% Minibatch loss at step 2000: 1.262677 Minibatch accuracy: 72.7% Validation accuracy: 77.9% Minibatch loss at step 2500: 0.777248 Minibatch accuracy: 83.6% Validation accuracy: 77.3% Minibatch loss at step 3000: 1.085464 Minibatch accuracy: 77.3% Validation accuracy: 78.6% Test accuracy: 86.1%
In [9]:
models.ix[0, 'graph'] = graph
print(models)
nRELUs accuracyVld accuracyTst \
0 0 78.57 86.098056
graph
0 <tensorflow.python.framework.ops.Graph object ...
Problem¶
Turn the logistic regression example with SGD into a 1-hidden layer neural network with rectified linear units (nn.relu()) and 1024 hidden nodes. This model should improve your validation / test accuracy_score.
In [10]:
nRELUs = [2 ** thsRelu for thsRelu in xrange(11)]
print(nRELUs)
for thsRelu in nRELUs:
models.ix[thsRelu, 'nRELUs'] = thsRelu
print models
[1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024]
nRELUs accuracyVld accuracyTst \
0 0 78.57 86.098056
1 1 NaN NaN
2 2 NaN NaN
4 4 NaN NaN
8 8 NaN NaN
16 16 NaN NaN
32 32 NaN NaN
64 64 NaN NaN
128 128 NaN NaN
256 256 NaN NaN
512 512 NaN NaN
1024 1024 NaN NaN
graph
0 <tensorflow.python.framework.ops.Graph object ...
1 NaN
2 NaN
4 NaN
8 NaN
16 NaN
32 NaN
64 NaN
128 NaN
256 NaN
512 NaN
1024 NaN
In [49]:
thsRelu = nRELUs[9]
batch_size = 128
graph = tf.Graph()
with graph.as_default():
# Input data. For the training data, we use a placeholder that will be fed
# at run time with a training minibatch.
tfwObsFitFtr = tf.placeholder(tf.float32,
shape=(batch_size, glbImg['size'] * glbImg['size']))
tfwObsFitRsp = tf.placeholder(tf.float32, shape=(batch_size, glbRspClassN))
tfwObsVldFtr = tf.constant(glbXVld)
tfwObsNewFtr = tf.constant(glbXNew)
# Variables.
tfwW1 = tf.Variable(
tf.truncated_normal([glbImg['size'] * glbImg['size'], thsRelu]), name = 'tfwW1')
tfwB1 = tf.Variable(tf.zeros([thsRelu]), name = 'tfwB1')
tfwW2 = tf.Variable(
tf.truncated_normal([thsRelu, glbRspClassN]), name = 'tfwW2')
tfwB2 = tf.Variable(tf.zeros([glbRspClassN]), name = 'tfwB2')
print(tfwW1.initialized_value())
print(tfwB1.initialized_value())
#print(relus.initialized_value())
print(tfwW2.initialized_value())
print(tfwB2.initialized_value())
#tf.Print(relus, [relus])
# Training computation.
layer1 = tf.matmul(tfwObsFitFtr, tfwW1) + tfwB1
layer2 = tf.nn.relu(layer1)
layer3 = tf.matmul(layer2, tfwW2) + tfwB2
loss = tf.reduce_mean(
tf.nn.softmax_cross_entropy_with_logits(layer3, tfwObsFitRsp))
# Optimizer.
optimizer = tf.train.GradientDescentOptimizer(0.5).minimize(loss)
# Predictions for the training, validation, and test data.
tfwObsTrnPred = tf.nn.softmax(layer3)
tfwObsVldPred = tf.nn.softmax(
tf.matmul(tf.nn.relu(tf.matmul(tfwObsVldFtr, tfwW1) + tfwB1), tfwW2) + tfwB2)
tfwObsNewPred = tf.nn.softmax(
tf.matmul(tf.nn.relu(tf.matmul(tfwObsNewFtr, tfwW1) + tfwB1), tfwW2) + tfwB2)
Tensor("Identity:0", shape=TensorShape([Dimension(784), Dimension(512)]), dtype=float32)
Tensor("Identity_1:0", shape=TensorShape([Dimension(512)]), dtype=float32)
Tensor("Identity_2:0", shape=TensorShape([Dimension(512), Dimension(10)]), dtype=float32)
Tensor("Identity_3:0", shape=TensorShape([Dimension(10)]), dtype=float32)
In [50]:
nStepsTfw = 3001
with tf.Session(graph=graph) as session:
tf.initialize_all_variables().run()
print("Initialized")
for step in range(nStepsTfw):
# Pick an offset within the training data, which has been randomized.
# Note: we could use better randomization across epochs.
offset = (step * batch_size) % (glbYFit.shape[0] - batch_size)
# Generate a minibatch.
batch_data = glbXFit[offset:(offset + batch_size), :]
batch_labels = glbYFit[offset:(offset + batch_size), :]
# Prepare a dictionary telling the session where to feed the minibatch.
# The key of the dictionary is the placeholder node of the graph to be fed,
# and the value is the numpy array to feed to it.
feed_dict = {tfwObsFitFtr : batch_data, tfwObsFitRsp : batch_labels}
_, l, predictions = session.run(
[optimizer, loss, tfwObsTrnPred], feed_dict=feed_dict)
if (step % 500 == 0):
print("Minibatch loss at step %d: %f" % (step, l))
print("Minibatch accuracy_score: %.1f%%" % accuracy_score(predictions, batch_labels))
print("Validation accuracy_score: %.1f%%" % accuracy_score(
tfwObsVldPred.eval(), glbYVld))
print("Test accuracy_score: %.1f%%" % accuracy_score(tfwObsNewPred.eval(), glbYNew))
models.ix[thsRelu, 'accuracy_scoreVld'] = accuracy_score(tfwObsVldPred.eval(), glbYVld)
models.ix[thsRelu, 'accuracy_scoreTst'] = accuracy_score( tfwObsNewPred.eval(), glbYNew)
models.ix[thsRelu, 'graph'] = graph
print(models)
Initialized
Minibatch loss at step 0: 235.251495
Minibatch accuracy: 14.1%
Validation accuracy: 23.9%
Minibatch loss at step 500: 15.635325
Minibatch accuracy: 72.7%
Validation accuracy: 77.7%
Minibatch loss at step 1000: 5.719280
Minibatch accuracy: 83.6%
Validation accuracy: 78.3%
Minibatch loss at step 1500: 3.931793
Minibatch accuracy: 76.6%
Validation accuracy: 75.8%
Minibatch loss at step 2000: 3.211185
Minibatch accuracy: 75.0%
Validation accuracy: 78.1%
Minibatch loss at step 2500: 1.988469
Minibatch accuracy: 80.5%
Validation accuracy: 78.0%
Minibatch loss at step 3000: 3.435107
Minibatch accuracy: 77.3%
Validation accuracy: 79.1%
Test accuracy: 86.7%
nRELUs accuracyVld accuracyTst \
0 0 78.57 86.098056
1 1 19.29 19.750053
2 2 36.26 40.087588
4 4 64.67 71.122623
8 8 75.80 82.861568
16 16 79.74 86.685537
32 32 76.77 84.287545
64 64 78.79 86.172826
128 128 78.86 86.210211
256 256 79.21 86.407819
512 512 79.12 86.674856
1024 1024 81.26 88.335826
graph
0 <tensorflow.python.framework.ops.Graph object ...
1 <tensorflow.python.framework.ops.Graph object ...
2 <tensorflow.python.framework.ops.Graph object ...
4 <tensorflow.python.framework.ops.Graph object ...
8 <tensorflow.python.framework.ops.Graph object ...
16 <tensorflow.python.framework.ops.Graph object ...
32 <tensorflow.python.framework.ops.Graph object ...
64 <tensorflow.python.framework.ops.Graph object ...
128 <tensorflow.python.framework.ops.Graph object ...
256 <tensorflow.python.framework.ops.Graph object ...
512 <tensorflow.python.framework.ops.Graph object ...
1024 <tensorflow.python.framework.ops.Graph object ...
In [52]:
plt.figure()
#plt.plot(models['nRELUs'], models['accuracy_score.fit'], 'bo-', label = 'fit')
plt.plot(models['nRELUs'], models['accuracy_scoreVld'], 'rs-', label = 'vld')
plt.plot(models['nRELUs'], models['accuracy_scoreTst'], 'gp-', label = 'new')
plt.legend(loc = 'lower right')
plt.title("accuracy_score")
plt.xscale('symlog', basex=2)
axes = plt.gca()
axes.set_xlabel('nRELUs')
# axes.set_xlim([mdlDF['l1_penalty'][mdlDF['RSS.vld'].argmin()] / 10 ** 2, \
# mdlDF['l1_penalty'][mdlDF['RSS.vld'].argmin()] * 10 ** 2])
# axes.set_ylim([0, mdlDF['RSS.vld'].min() * 1.5])
plt.show()
In [ ]: